{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 37.45485154576064, "eval_steps": 500, "global_step": 61200, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.002298680927682047, "clip_ratio/high_mean": 0.0014097090575281375, "clip_ratio/low_mean": 0.0010746494022896512, "clip_ratio/low_min": 0.000491498626312629, "clip_ratio/region_mean": 0.0024843584857687042, "completions/clipped_ratio": 0.0, "completions/max_length": 628.9, "completions/max_terminated_length": 628.9, "completions/mean_length": 192.80703430175782, "completions/mean_terminated_length": 192.80703430175782, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03368704972478251, "epoch": 0.018365472910927456, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.005279541015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3040539.0, "reward": 7.9436064720153805, "reward_std": 0.3478423520922661, "rewards/combined_reward/mean": 7.94360613822937, "rewards/combined_reward/std": 2.310935604572296, "step": 30 }, { "clip_ratio/high_max": 0.002078863222656461, "clip_ratio/high_mean": 0.0012842626677108152, "clip_ratio/low_mean": 0.0007736816845863359, "clip_ratio/low_min": 0.00019135953067840698, "clip_ratio/region_mean": 0.0020579443439298, "completions/clipped_ratio": 0.0, "completions/max_length": 855.7, "completions/max_terminated_length": 855.7, "completions/mean_length": 231.6408889770508, "completions/mean_terminated_length": 231.6408889770508, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.03593083626280228, "epoch": 0.03673094582185491, "frac_reward_zero_std": 0.11666667126119137, "grad_norm": 0.032470703125, "learning_rate": 5e-06, "loss": 0.0029, "num_tokens": 6230200.0, "reward": 8.096699523925782, "reward_std": 0.4153560519218445, "rewards/combined_reward/mean": 8.096699142456055, "rewards/combined_reward/std": 2.33445348739624, "step": 60 }, { "clip_ratio/high_max": 0.0022936015457768614, "clip_ratio/high_mean": 0.0012406377569883869, "clip_ratio/low_mean": 0.0007651677459458976, "clip_ratio/low_min": 0.00018769118905765937, "clip_ratio/region_mean": 0.0020058055233675987, "completions/clipped_ratio": 0.0, "completions/max_length": 578.7, "completions/max_terminated_length": 578.7, "completions/mean_length": 188.04922332763672, "completions/mean_terminated_length": 188.04922332763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.0310395099921152, "epoch": 0.05509641873278237, "frac_reward_zero_std": 0.23333334028720856, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 9252469.0, "reward": 7.911698579788208, "reward_std": 0.4084559828042984, "rewards/combined_reward/mean": 7.911698627471924, "rewards/combined_reward/std": 2.5382960200309754, "step": 90 }, { "clip_ratio/high_max": 0.001899963257407459, "clip_ratio/high_mean": 0.0010795801771261418, "clip_ratio/low_mean": 0.0007598962124272172, "clip_ratio/low_min": 0.00024801954496069813, "clip_ratio/region_mean": 0.001839476383368795, "completions/clipped_ratio": 0.0, "completions/max_length": 609.4, "completions/max_terminated_length": 609.4, "completions/mean_length": 193.97891235351562, "completions/mean_terminated_length": 193.97891235351562, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03328208299353719, "epoch": 0.07346189164370982, "frac_reward_zero_std": 0.15833334028720855, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 12297508.0, "reward": 7.9483390808105465, "reward_std": 0.41580687314271925, "rewards/combined_reward/mean": 7.948338603973388, "rewards/combined_reward/std": 2.351696050167084, "step": 120 }, { "clip_ratio/high_max": 0.002070214704144746, "clip_ratio/high_mean": 0.0011314062478049892, "clip_ratio/low_mean": 0.0008766344949738898, "clip_ratio/low_min": 0.0002453822829314352, "clip_ratio/region_mean": 0.0020080407334414, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 198.45938110351562, "completions/mean_terminated_length": 198.45938110351562, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.030770368625720343, "epoch": 0.09182736455463728, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 15359752.0, "reward": 7.92764253616333, "reward_std": 0.3968218892812729, "rewards/combined_reward/mean": 7.927642202377319, "rewards/combined_reward/std": 2.3746426463127137, "step": 150 }, { "clip_ratio/high_max": 0.0018164947085703412, "clip_ratio/high_mean": 0.0010177146095278053, "clip_ratio/low_mean": 0.0008935027972862979, "clip_ratio/low_min": 0.0002359225230369096, "clip_ratio/region_mean": 0.001911217401114603, "completions/clipped_ratio": 0.0, "completions/max_length": 690.7, "completions/max_terminated_length": 690.7, "completions/mean_length": 197.6932342529297, "completions/mean_terminated_length": 197.6932342529297, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.029749292197326818, "epoch": 0.11019283746556474, "frac_reward_zero_std": 0.16250000670552253, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 18419054.0, "reward": 7.699309015274048, "reward_std": 0.37132823169231416, "rewards/combined_reward/mean": 7.699308824539185, "rewards/combined_reward/std": 2.6705421686172484, "step": 180 }, { "clip_ratio/high_max": 0.0024637741856471016, "clip_ratio/high_mean": 0.0014221442819689401, "clip_ratio/low_mean": 0.000910129641185146, "clip_ratio/low_min": 0.0003086385899223387, "clip_ratio/region_mean": 0.002332273938615496, "completions/clipped_ratio": 0.0, "completions/max_length": 694.7, "completions/max_terminated_length": 694.7, "completions/mean_length": 211.04479522705077, "completions/mean_terminated_length": 211.04479522705077, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03607216991173724, "epoch": 0.1285583103764922, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 21529626.0, "reward": 8.00140290260315, "reward_std": 0.44595648944377897, "rewards/combined_reward/mean": 8.001402664184571, "rewards/combined_reward/std": 2.3186160564422607, "step": 210 }, { "clip_ratio/high_max": 0.0024167197315061156, "clip_ratio/high_mean": 0.0013561515722055144, "clip_ratio/low_mean": 0.0007502043120136174, "clip_ratio/low_min": 0.00020040508000723395, "clip_ratio/region_mean": 0.002106355910655111, "completions/clipped_ratio": 0.0, "completions/max_length": 569.4, "completions/max_terminated_length": 569.4, "completions/mean_length": 193.96927490234376, "completions/mean_terminated_length": 193.96927490234376, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03483728159529467, "epoch": 0.14692378328741965, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 24574628.0, "reward": 8.014149522781372, "reward_std": 0.44810320287942884, "rewards/combined_reward/mean": 8.014149236679078, "rewards/combined_reward/std": 2.1841938972473143, "step": 240 }, { "clip_ratio/high_max": 0.001968248602740156, "clip_ratio/high_mean": 0.0011964700664975681, "clip_ratio/low_mean": 0.0008757076672433565, "clip_ratio/low_min": 0.00024700483190827073, "clip_ratio/region_mean": 0.002072177730345478, "completions/clipped_ratio": 0.0, "completions/max_length": 634.2, "completions/max_terminated_length": 634.2, "completions/mean_length": 203.79974212646485, "completions/mean_terminated_length": 203.79974212646485, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03070779498666525, "epoch": 0.1652892561983471, "frac_reward_zero_std": 0.15416666828095912, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 27657379.0, "reward": 7.97825403213501, "reward_std": 0.35415989756584165, "rewards/combined_reward/mean": 7.9782538414001465, "rewards/combined_reward/std": 2.484586763381958, "step": 270 }, { "clip_ratio/high_max": 0.0015550208484152487, "clip_ratio/high_mean": 0.0009030932484165533, "clip_ratio/low_mean": 0.0008360420658694542, "clip_ratio/low_min": 0.00024489574279868976, "clip_ratio/region_mean": 0.0017391353258669065, "completions/clipped_ratio": 0.0, "completions/max_length": 596.3, "completions/max_terminated_length": 596.3, "completions/mean_length": 185.0432327270508, "completions/mean_terminated_length": 185.0432327270508, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02996512919974824, "epoch": 0.18365472910927455, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 30668105.0, "reward": 8.006351327896118, "reward_std": 0.3742027297616005, "rewards/combined_reward/mean": 8.006350994110107, "rewards/combined_reward/std": 2.36344553232193, "step": 300 }, { "clip_ratio/high_max": 0.0025060093840390133, "clip_ratio/high_mean": 0.001400553419322629, "clip_ratio/low_mean": 0.0008873143381303331, "clip_ratio/low_min": 0.0002824767114361748, "clip_ratio/region_mean": 0.0022878677322296427, "completions/clipped_ratio": 0.0, "completions/max_length": 593.3, "completions/max_terminated_length": 593.3, "completions/mean_length": 199.05937957763672, "completions/mean_terminated_length": 199.05937957763672, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03429864391994973, "epoch": 0.20202020202020202, "frac_reward_zero_std": 0.18333333693444728, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0031, "num_tokens": 33732653.0, "reward": 7.775531387329101, "reward_std": 0.4567530244588852, "rewards/combined_reward/mean": 7.775531244277954, "rewards/combined_reward/std": 2.575367820262909, "step": 330 }, { "clip_ratio/high_max": 0.002146991718230614, "clip_ratio/high_mean": 0.0012505734921433032, "clip_ratio/low_mean": 0.0008999307300352181, "clip_ratio/low_min": 0.0003246314954594709, "clip_ratio/region_mean": 0.002150504228969415, "completions/clipped_ratio": 0.0, "completions/max_length": 571.7, "completions/max_terminated_length": 571.7, "completions/mean_length": 205.3429718017578, "completions/mean_terminated_length": 205.3429718017578, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.029647630111624797, "epoch": 0.22038567493112948, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 36821330.0, "reward": 7.961780118942261, "reward_std": 0.387213459610939, "rewards/combined_reward/mean": 7.96177978515625, "rewards/combined_reward/std": 2.3776774644851684, "step": 360 }, { "clip_ratio/high_max": 0.0020972756455497196, "clip_ratio/high_mean": 0.0011828351930792754, "clip_ratio/low_mean": 0.0009306608998789063, "clip_ratio/low_min": 0.0003554193375748582, "clip_ratio/region_mean": 0.0021134960964748948, "completions/clipped_ratio": 0.0, "completions/max_length": 611.2, "completions/max_terminated_length": 611.2, "completions/mean_length": 197.77969207763672, "completions/mean_terminated_length": 197.77969207763672, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03194400953749816, "epoch": 0.23875114784205692, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 39880964.0, "reward": 8.048375511169434, "reward_std": 0.4133232310414314, "rewards/combined_reward/mean": 8.04837508201599, "rewards/combined_reward/std": 2.149471378326416, "step": 390 }, { "clip_ratio/high_max": 0.002085476055314454, "clip_ratio/high_mean": 0.0011323032044553353, "clip_ratio/low_mean": 0.0008798981811802757, "clip_ratio/low_min": 0.00027899231160214794, "clip_ratio/region_mean": 0.0020122014017639835, "completions/clipped_ratio": 0.0, "completions/max_length": 533.8, "completions/max_terminated_length": 533.8, "completions/mean_length": 185.28411865234375, "completions/mean_terminated_length": 185.28411865234375, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03396148599373797, "epoch": 0.2571166207529844, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 42892615.0, "reward": 7.931128358840942, "reward_std": 0.4712811321020126, "rewards/combined_reward/mean": 7.931128311157226, "rewards/combined_reward/std": 2.3486614346504213, "step": 420 }, { "clip_ratio/high_max": 0.002185002150266276, "clip_ratio/high_mean": 0.0012570201739436015, "clip_ratio/low_mean": 0.0008240062127394291, "clip_ratio/low_min": 0.0002290480855056861, "clip_ratio/region_mean": 0.002081026394929116, "completions/clipped_ratio": 0.0, "completions/max_length": 552.3, "completions/max_terminated_length": 552.3, "completions/mean_length": 207.06302642822266, "completions/mean_terminated_length": 207.06302642822266, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.031750526581890884, "epoch": 0.27548209366391185, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 45987897.0, "reward": 8.026388454437257, "reward_std": 0.4225667893886566, "rewards/combined_reward/mean": 8.02638840675354, "rewards/combined_reward/std": 2.275655543804169, "step": 450 }, { "clip_ratio/high_max": 0.001923265476943925, "clip_ratio/high_mean": 0.0011861504157423043, "clip_ratio/low_mean": 0.0009366966602101457, "clip_ratio/low_min": 0.00028121883418255795, "clip_ratio/region_mean": 0.0021228470858962586, "completions/clipped_ratio": 0.0, "completions/max_length": 618.2, "completions/max_terminated_length": 618.2, "completions/mean_length": 206.7869842529297, "completions/mean_terminated_length": 206.7869842529297, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.03517654212191701, "epoch": 0.2938475665748393, "frac_reward_zero_std": 0.17916667200624942, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0031, "num_tokens": 49082119.0, "reward": 7.991330814361572, "reward_std": 0.41407146453857424, "rewards/combined_reward/mean": 7.991330528259278, "rewards/combined_reward/std": 2.2004170536994936, "step": 480 }, { "clip_ratio/high_max": 0.0016832665860420093, "clip_ratio/high_mean": 0.0009793541384472823, "clip_ratio/low_mean": 0.000746260083966869, "clip_ratio/low_min": 0.00018863383011193945, "clip_ratio/region_mean": 0.001725614244545189, "completions/clipped_ratio": 0.0, "completions/max_length": 583.7, "completions/max_terminated_length": 583.7, "completions/mean_length": 203.46380767822265, "completions/mean_terminated_length": 203.46380767822265, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02917417084487776, "epoch": 0.3122130394857668, "frac_reward_zero_std": 0.15000000260770321, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 52163580.0, "reward": 7.8119344234466555, "reward_std": 0.3493504047393799, "rewards/combined_reward/mean": 7.811933946609497, "rewards/combined_reward/std": 2.422358202934265, "step": 510 }, { "clip_ratio/high_max": 0.002217247691199494, "clip_ratio/high_mean": 0.0013235722920702149, "clip_ratio/low_mean": 0.000755532556407464, "clip_ratio/low_min": 0.0002267678357990614, "clip_ratio/region_mean": 0.0020791048290751254, "completions/clipped_ratio": 0.0, "completions/max_length": 622.4, "completions/max_terminated_length": 622.4, "completions/mean_length": 199.71536865234376, "completions/mean_terminated_length": 199.71536865234376, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03585367669972281, "epoch": 0.3305785123966942, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 55230647.0, "reward": 7.939273691177368, "reward_std": 0.3796018198132515, "rewards/combined_reward/mean": 7.93927321434021, "rewards/combined_reward/std": 2.369417893886566, "step": 540 }, { "clip_ratio/high_max": 0.002426819514221279, "clip_ratio/high_mean": 0.0012794760583346942, "clip_ratio/low_mean": 0.0010436177211280058, "clip_ratio/low_min": 0.0002597648211425015, "clip_ratio/region_mean": 0.0023230937620004016, "completions/clipped_ratio": 0.0, "completions/max_length": 593.2, "completions/max_terminated_length": 593.2, "completions/mean_length": 205.60365295410156, "completions/mean_terminated_length": 205.60365295410156, "completions/min_length": 46.5, "completions/min_terminated_length": 46.5, "entropy": 0.03566559550041954, "epoch": 0.34894398530762166, "frac_reward_zero_std": 0.11666666828095913, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 58320325.0, "reward": 8.099483823776245, "reward_std": 0.35778065025806427, "rewards/combined_reward/mean": 8.099483346939087, "rewards/combined_reward/std": 2.148952579498291, "step": 570 }, { "clip_ratio/high_max": 0.0021363691354054025, "clip_ratio/high_mean": 0.0011542135917504007, "clip_ratio/low_mean": 0.001103763699575211, "clip_ratio/low_min": 0.0004278326297935564, "clip_ratio/region_mean": 0.002257977293629665, "completions/clipped_ratio": 0.0, "completions/max_length": 546.6, "completions/max_terminated_length": 546.6, "completions/mean_length": 201.9737060546875, "completions/mean_terminated_length": 201.9737060546875, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.03500958999308447, "epoch": 0.3673094582185491, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 61396064.0, "reward": 8.016637134552003, "reward_std": 0.38001810014247894, "rewards/combined_reward/mean": 8.01663670539856, "rewards/combined_reward/std": 2.4362794756889343, "step": 600 }, { "clip_ratio/high_max": 0.002214815512706991, "clip_ratio/high_mean": 0.001350016344076721, "clip_ratio/low_mean": 0.0009054154167339826, "clip_ratio/low_min": 0.00031790575327856154, "clip_ratio/region_mean": 0.0022554317557175333, "completions/clipped_ratio": 0.0, "completions/max_length": 594.6, "completions/max_terminated_length": 594.6, "completions/mean_length": 212.28203735351562, "completions/mean_terminated_length": 212.28203735351562, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.03339529680088162, "epoch": 0.3856749311294766, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 64511387.0, "reward": 7.7785224437713625, "reward_std": 0.45537002831697465, "rewards/combined_reward/mean": 7.778522300720215, "rewards/combined_reward/std": 2.519152307510376, "step": 630 }, { "clip_ratio/high_max": 0.0021392674442419473, "clip_ratio/high_mean": 0.0012687505547849772, "clip_ratio/low_mean": 0.000890311476541683, "clip_ratio/low_min": 0.0002834693225546895, "clip_ratio/region_mean": 0.002159062038602618, "completions/clipped_ratio": 0.0, "completions/max_length": 568.6, "completions/max_terminated_length": 568.6, "completions/mean_length": 190.4362014770508, "completions/mean_terminated_length": 190.4362014770508, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.031103672040626407, "epoch": 0.40404040404040403, "frac_reward_zero_std": 0.1625000022351742, "grad_norm": 0.01239013671875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 67542822.0, "reward": 8.137095832824707, "reward_std": 0.36042298972606657, "rewards/combined_reward/mean": 8.13709545135498, "rewards/combined_reward/std": 2.081758749485016, "step": 660 }, { "clip_ratio/high_max": 0.0022281054305494765, "clip_ratio/high_mean": 0.001321178903648009, "clip_ratio/low_mean": 0.0009817815400310792, "clip_ratio/low_min": 0.0003020630775912044, "clip_ratio/region_mean": 0.0023029604208810875, "completions/clipped_ratio": 0.0, "completions/max_length": 608.6, "completions/max_terminated_length": 608.6, "completions/mean_length": 198.03802795410155, "completions/mean_terminated_length": 198.03802795410155, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03624664616460602, "epoch": 0.42240587695133147, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 70603448.0, "reward": 7.912246036529541, "reward_std": 0.4260909706354141, "rewards/combined_reward/mean": 7.912245798110962, "rewards/combined_reward/std": 2.5427329540252686, "step": 690 }, { "clip_ratio/high_max": 0.0020439808688630973, "clip_ratio/high_mean": 0.0012038245557050686, "clip_ratio/low_mean": 0.0008579153529960119, "clip_ratio/low_min": 0.00024223218909658802, "clip_ratio/region_mean": 0.0020617399025165167, "completions/clipped_ratio": 0.0, "completions/max_length": 545.9, "completions/max_terminated_length": 545.9, "completions/mean_length": 193.91614990234376, "completions/mean_terminated_length": 193.91614990234376, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03323651060151557, "epoch": 0.44077134986225897, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 73648246.0, "reward": 7.815176105499267, "reward_std": 0.38900420963764193, "rewards/combined_reward/mean": 7.8151758193969725, "rewards/combined_reward/std": 2.547624635696411, "step": 720 }, { "clip_ratio/high_max": 0.0020780818854594448, "clip_ratio/high_mean": 0.0012070366632542574, "clip_ratio/low_mean": 0.0008821424686175306, "clip_ratio/low_min": 0.00027388027616931745, "clip_ratio/region_mean": 0.002089179110286447, "completions/clipped_ratio": 0.0, "completions/max_length": 620.3, "completions/max_terminated_length": 620.3, "completions/mean_length": 199.87891235351563, "completions/mean_terminated_length": 199.87891235351563, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.030513445955390732, "epoch": 0.4591368227731864, "frac_reward_zero_std": 0.1291666716337204, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 76715941.0, "reward": 7.926436948776245, "reward_std": 0.44384547621011733, "rewards/combined_reward/mean": 7.926436471939087, "rewards/combined_reward/std": 2.1829285740852358, "step": 750 }, { "clip_ratio/high_max": 0.001910028778850877, "clip_ratio/high_mean": 0.001148136201906406, "clip_ratio/low_mean": 0.0009527590798825258, "clip_ratio/low_min": 0.0003183514165963667, "clip_ratio/region_mean": 0.002100895287973496, "completions/clipped_ratio": 0.0, "completions/max_length": 667.3, "completions/max_terminated_length": 667.3, "completions/mean_length": 198.6487014770508, "completions/mean_terminated_length": 198.6487014770508, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.0312024826494356, "epoch": 0.47750229568411384, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 79778912.0, "reward": 8.07279577255249, "reward_std": 0.3536663591861725, "rewards/combined_reward/mean": 8.072795343399047, "rewards/combined_reward/std": 2.2697670578956606, "step": 780 }, { "clip_ratio/high_max": 0.0021325662077288144, "clip_ratio/high_mean": 0.0012693239993192644, "clip_ratio/low_mean": 0.0010402008418168406, "clip_ratio/low_min": 0.00028348741010025456, "clip_ratio/region_mean": 0.0023095248485333285, "completions/clipped_ratio": 0.0, "completions/max_length": 540.1, "completions/max_terminated_length": 540.1, "completions/mean_length": 197.88229370117188, "completions/mean_terminated_length": 197.88229370117188, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.03432531516688565, "epoch": 0.49586776859504134, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 82838940.0, "reward": 7.948207044601441, "reward_std": 0.37523208409547804, "rewards/combined_reward/mean": 7.9482067108154295, "rewards/combined_reward/std": 2.5029421210289002, "step": 810 }, { "clip_ratio/high_max": 0.0017835694675644239, "clip_ratio/high_mean": 0.0010061166416562629, "clip_ratio/low_mean": 0.0008966126165129633, "clip_ratio/low_min": 0.0002476482327134969, "clip_ratio/region_mean": 0.001902729267506705, "completions/clipped_ratio": 0.0, "completions/max_length": 597.8, "completions/max_terminated_length": 597.8, "completions/mean_length": 194.90052642822266, "completions/mean_terminated_length": 194.90052642822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029072246816940605, "epoch": 0.5142332415059688, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.009765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 85887518.0, "reward": 7.793392181396484, "reward_std": 0.3869756281375885, "rewards/combined_reward/mean": 7.793391942977905, "rewards/combined_reward/std": 2.5091043591499327, "step": 840 }, { "clip_ratio/high_max": 0.002120046388396683, "clip_ratio/high_mean": 0.001207983260974288, "clip_ratio/low_mean": 0.0008034145355244012, "clip_ratio/low_min": 0.00017733471904648467, "clip_ratio/region_mean": 0.0020113977884951358, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 190.8268295288086, "completions/mean_terminated_length": 190.8268295288086, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.0315463552561899, "epoch": 0.5325987144168962, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 88920453.0, "reward": 7.834636545181274, "reward_std": 0.2789326697587967, "rewards/combined_reward/mean": 7.834636163711548, "rewards/combined_reward/std": 2.4560495972633363, "step": 870 }, { "clip_ratio/high_max": 0.0016731901036109776, "clip_ratio/high_mean": 0.0009879581821830167, "clip_ratio/low_mean": 0.0007805553903987554, "clip_ratio/low_min": 0.00020795927751654137, "clip_ratio/region_mean": 0.0017685135831319106, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 198.26016082763672, "completions/mean_terminated_length": 198.26016082763672, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02851883862167597, "epoch": 0.5509641873278237, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.004425048828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 91981932.0, "reward": 8.30056552886963, "reward_std": 0.2963122457265854, "rewards/combined_reward/mean": 8.300565052032471, "rewards/combined_reward/std": 2.106830906867981, "step": 900 }, { "clip_ratio/high_max": 0.0019623216435623665, "clip_ratio/high_mean": 0.0010570029233349487, "clip_ratio/low_mean": 0.0008877858126652427, "clip_ratio/low_min": 0.00030979815710452385, "clip_ratio/region_mean": 0.0019447887529774259, "completions/clipped_ratio": 0.0, "completions/max_length": 615.8, "completions/max_terminated_length": 615.8, "completions/mean_length": 201.97760925292968, "completions/mean_terminated_length": 201.97760925292968, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.032136041484773156, "epoch": 0.5693296602387512, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 95057686.0, "reward": 7.646303367614746, "reward_std": 0.3974746435880661, "rewards/combined_reward/mean": 7.646303033828735, "rewards/combined_reward/std": 2.5222560286521913, "step": 930 }, { "clip_ratio/high_max": 0.0019516731743351556, "clip_ratio/high_mean": 0.0011493602527480108, "clip_ratio/low_mean": 0.0009509573840963034, "clip_ratio/low_min": 0.00032564764381580366, "clip_ratio/region_mean": 0.0021003176210797393, "completions/clipped_ratio": 0.0, "completions/max_length": 601.8, "completions/max_terminated_length": 601.8, "completions/mean_length": 216.67005767822266, "completions/mean_terminated_length": 216.67005767822266, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 0.030612992231423657, "epoch": 0.5876951331496786, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 98189859.0, "reward": 7.90199990272522, "reward_std": 0.3943075627088547, "rewards/combined_reward/mean": 7.9019999504089355, "rewards/combined_reward/std": 2.3029553174972532, "step": 960 }, { "clip_ratio/high_max": 0.0017944580918992869, "clip_ratio/high_mean": 0.0010733590150872865, "clip_ratio/low_mean": 0.0008440976239702043, "clip_ratio/low_min": 0.00024060041517562543, "clip_ratio/region_mean": 0.0019174566521542146, "completions/clipped_ratio": 0.0, "completions/max_length": 582.6, "completions/max_terminated_length": 582.6, "completions/mean_length": 207.4651092529297, "completions/mean_terminated_length": 207.4651092529297, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.03010669330445429, "epoch": 0.6060606060606061, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.003, "num_tokens": 101286685.0, "reward": 8.056331729888916, "reward_std": 0.4333717182278633, "rewards/combined_reward/mean": 8.056331539154053, "rewards/combined_reward/std": 2.1923380494117737, "step": 990 }, { "clip_ratio/high_max": 0.0018877545420158033, "clip_ratio/high_mean": 0.001069604901689066, "clip_ratio/low_mean": 0.0007127766915800748, "clip_ratio/low_min": 0.00020388083042538103, "clip_ratio/region_mean": 0.0017823815857506513, "completions/clipped_ratio": 0.0, "completions/max_length": 593.9, "completions/max_terminated_length": 593.9, "completions/mean_length": 198.8546936035156, "completions/mean_terminated_length": 198.8546936035156, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.029684732722428938, "epoch": 0.6244260789715336, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 104350447.0, "reward": 8.14956169128418, "reward_std": 0.33822038769721985, "rewards/combined_reward/mean": 8.149561643600464, "rewards/combined_reward/std": 2.197388803958893, "step": 1020 }, { "clip_ratio/high_max": 0.0023587385929810505, "clip_ratio/high_mean": 0.001342645982125153, "clip_ratio/low_mean": 0.0008290950462348216, "clip_ratio/low_min": 0.00022354230459313839, "clip_ratio/region_mean": 0.002171741030178964, "completions/clipped_ratio": 0.0, "completions/max_length": 576.3, "completions/max_terminated_length": 576.3, "completions/mean_length": 210.5812561035156, "completions/mean_terminated_length": 210.5812561035156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.033682461362332106, "epoch": 0.642791551882461, "frac_reward_zero_std": 0.1166666705161333, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 107459239.0, "reward": 8.043930768966675, "reward_std": 0.4170063197612762, "rewards/combined_reward/mean": 8.043930768966675, "rewards/combined_reward/std": 2.1446541786193847, "step": 1050 }, { "clip_ratio/high_max": 0.0022498432551704656, "clip_ratio/high_mean": 0.0012173254324200874, "clip_ratio/low_mean": 0.0006536847591632977, "clip_ratio/low_min": 0.00018302236809783305, "clip_ratio/region_mean": 0.0018710102119560664, "completions/clipped_ratio": 0.0, "completions/max_length": 539.9, "completions/max_terminated_length": 539.9, "completions/mean_length": 199.14297180175782, "completions/mean_terminated_length": 199.14297180175782, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.030886966657514375, "epoch": 0.6611570247933884, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 110524108.0, "reward": 8.030812501907349, "reward_std": 0.36075071394443514, "rewards/combined_reward/mean": 8.030812168121338, "rewards/combined_reward/std": 2.2879400730133055, "step": 1080 }, { "clip_ratio/high_max": 0.00228227252518991, "clip_ratio/high_mean": 0.0012713794617108458, "clip_ratio/low_mean": 0.0009722242016626599, "clip_ratio/low_min": 0.0002986928947696773, "clip_ratio/region_mean": 0.0022436036844737826, "completions/clipped_ratio": 0.0, "completions/max_length": 569.8, "completions/max_terminated_length": 569.8, "completions/mean_length": 201.97448272705077, "completions/mean_terminated_length": 201.97448272705077, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03322044846136123, "epoch": 0.6795224977043158, "frac_reward_zero_std": 0.17083333544433116, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0029, "num_tokens": 113599850.0, "reward": 7.8216651439666744, "reward_std": 0.41930296570062636, "rewards/combined_reward/mean": 7.8216650009155275, "rewards/combined_reward/std": 2.3604644417762755, "step": 1110 }, { "clip_ratio/high_max": 0.0021269196634724117, "clip_ratio/high_mean": 0.001164386381181733, "clip_ratio/low_mean": 0.0008923814615021305, "clip_ratio/low_min": 0.00024172754007546852, "clip_ratio/region_mean": 0.0020567678341952464, "completions/clipped_ratio": 0.0, "completions/max_length": 526.9, "completions/max_terminated_length": 526.9, "completions/mean_length": 202.31406860351564, "completions/mean_terminated_length": 202.31406860351564, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.030594368015105525, "epoch": 0.6978879706152433, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 116676896.0, "reward": 8.070734786987305, "reward_std": 0.3391618192195892, "rewards/combined_reward/mean": 8.070734643936158, "rewards/combined_reward/std": 2.148129713535309, "step": 1140 }, { "clip_ratio/high_max": 0.0020913353553623894, "clip_ratio/high_mean": 0.0011463178610332154, "clip_ratio/low_mean": 0.0009352680429098351, "clip_ratio/low_min": 0.00034068675498322897, "clip_ratio/region_mean": 0.0020815859010326677, "completions/clipped_ratio": 0.0, "completions/max_length": 580.7, "completions/max_terminated_length": 580.7, "completions/mean_length": 214.31771240234374, "completions/mean_terminated_length": 214.31771240234374, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.032632776986186704, "epoch": 0.7162534435261708, "frac_reward_zero_std": 0.11250000298023224, "grad_norm": 0.00421142578125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 119800036.0, "reward": 8.284668064117431, "reward_std": 0.35448867082595825, "rewards/combined_reward/mean": 8.284667539596558, "rewards/combined_reward/std": 1.9649291515350342, "step": 1170 }, { "clip_ratio/high_max": 0.002154639967678425, "clip_ratio/high_mean": 0.001268926506357578, "clip_ratio/low_mean": 0.0008197877277173877, "clip_ratio/low_min": 0.000275555020198226, "clip_ratio/region_mean": 0.0020887142352876255, "completions/clipped_ratio": 0.0, "completions/max_length": 570.1, "completions/max_terminated_length": 570.1, "completions/mean_length": 206.11042175292968, "completions/mean_terminated_length": 206.11042175292968, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.0313661210394154, "epoch": 0.7346189164370982, "frac_reward_zero_std": 0.18333333544433117, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 122891660.0, "reward": 7.837224817276001, "reward_std": 0.3958782717585564, "rewards/combined_reward/mean": 7.837224721908569, "rewards/combined_reward/std": 2.6212992668151855, "step": 1200 }, { "clip_ratio/high_max": 0.0017681092038401402, "clip_ratio/high_mean": 0.0010538797259262841, "clip_ratio/low_mean": 0.0007117482352138419, "clip_ratio/low_min": 0.00020373716809748052, "clip_ratio/region_mean": 0.0017656279548342961, "completions/clipped_ratio": 0.0, "completions/max_length": 651.9, "completions/max_terminated_length": 651.9, "completions/mean_length": 208.9468795776367, "completions/mean_terminated_length": 208.9468795776367, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.028569689812138678, "epoch": 0.7529843893480257, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 125994176.0, "reward": 8.245858478546143, "reward_std": 0.34928580820560456, "rewards/combined_reward/mean": 8.245858192443848, "rewards/combined_reward/std": 2.0632929503917694, "step": 1230 }, { "clip_ratio/high_max": 0.0022446903128487366, "clip_ratio/high_mean": 0.0013506480481737526, "clip_ratio/low_mean": 0.0007966001337384417, "clip_ratio/low_min": 0.0002643085046050449, "clip_ratio/region_mean": 0.0021472482151390676, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 208.2635467529297, "completions/mean_terminated_length": 208.2635467529297, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.03343908980799218, "epoch": 0.7713498622589532, "frac_reward_zero_std": 0.15416667349636554, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 129094068.0, "reward": 8.016087436676026, "reward_std": 0.4137433461844921, "rewards/combined_reward/mean": 8.016087198257447, "rewards/combined_reward/std": 2.2964460968971254, "step": 1260 }, { "clip_ratio/high_max": 0.002129952279938152, "clip_ratio/high_mean": 0.0011449313540651928, "clip_ratio/low_mean": 0.0008588203973583101, "clip_ratio/low_min": 0.00028579624825700496, "clip_ratio/region_mean": 0.002003751753848822, "completions/clipped_ratio": 0.0, "completions/max_length": 542.7, "completions/max_terminated_length": 542.7, "completions/mean_length": 205.31302795410156, "completions/mean_terminated_length": 205.31302795410156, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.03142261994071305, "epoch": 0.7897153351698806, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 132182630.0, "reward": 8.106445074081421, "reward_std": 0.2894119918346405, "rewards/combined_reward/mean": 8.106444835662842, "rewards/combined_reward/std": 2.248934280872345, "step": 1290 }, { "clip_ratio/high_max": 0.0021709228528682917, "clip_ratio/high_mean": 0.0012978017525407874, "clip_ratio/low_mean": 0.0008506177954889911, "clip_ratio/low_min": 0.0002333250071387738, "clip_ratio/region_mean": 0.0021484195206236717, "completions/clipped_ratio": 0.0, "completions/max_length": 695.4, "completions/max_terminated_length": 695.4, "completions/mean_length": 210.56849517822266, "completions/mean_terminated_length": 210.56849517822266, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03550994681815307, "epoch": 0.8080808080808081, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.01239013671875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 135291373.0, "reward": 8.045622491836548, "reward_std": 0.349932986497879, "rewards/combined_reward/mean": 8.0456223487854, "rewards/combined_reward/std": 2.15114803314209, "step": 1320 }, { "clip_ratio/high_max": 0.0021224123192951083, "clip_ratio/high_mean": 0.0012427997489188177, "clip_ratio/low_mean": 0.0008560406058677472, "clip_ratio/low_min": 0.00022852918264106847, "clip_ratio/region_mean": 0.002098840347995671, "completions/clipped_ratio": 0.0, "completions/max_length": 550.8, "completions/max_terminated_length": 550.8, "completions/mean_length": 195.23099212646486, "completions/mean_terminated_length": 195.23099212646486, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03145919463907679, "epoch": 0.8264462809917356, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 138341220.0, "reward": 7.870165491104126, "reward_std": 0.37761526852846145, "rewards/combined_reward/mean": 7.870165061950684, "rewards/combined_reward/std": 2.567016911506653, "step": 1350 }, { "clip_ratio/high_max": 0.0024041583710641135, "clip_ratio/high_mean": 0.0012989246838211936, "clip_ratio/low_mean": 0.0009309142621835539, "clip_ratio/low_min": 0.0003316029367852025, "clip_ratio/region_mean": 0.002229838953159439, "completions/clipped_ratio": 0.0, "completions/max_length": 626.8, "completions/max_terminated_length": 626.8, "completions/mean_length": 202.8377655029297, "completions/mean_terminated_length": 202.8377655029297, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.03579080277122557, "epoch": 0.8448117539026629, "frac_reward_zero_std": 0.1958333406597376, "grad_norm": 0.004302978515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 141420277.0, "reward": 8.213984441757201, "reward_std": 0.38256327956914904, "rewards/combined_reward/mean": 8.213984251022339, "rewards/combined_reward/std": 2.016675078868866, "step": 1380 }, { "clip_ratio/high_max": 0.0022367905772019488, "clip_ratio/high_mean": 0.0012412943389790599, "clip_ratio/low_mean": 0.0008453275833138226, "clip_ratio/low_min": 0.0002259812155292214, "clip_ratio/region_mean": 0.0020866218760299186, "completions/clipped_ratio": 0.0, "completions/max_length": 586.4, "completions/max_terminated_length": 586.4, "completions/mean_length": 201.46953582763672, "completions/mean_terminated_length": 201.46953582763672, "completions/min_length": 46.9, "completions/min_terminated_length": 46.9, "entropy": 0.03081163459767898, "epoch": 0.8631772268135904, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 144494080.0, "reward": 8.364498186111451, "reward_std": 0.31161497682332995, "rewards/combined_reward/mean": 8.364497995376587, "rewards/combined_reward/std": 1.932727962732315, "step": 1410 }, { "clip_ratio/high_max": 0.0023449233451780555, "clip_ratio/high_mean": 0.0013816186646484615, "clip_ratio/low_mean": 0.0007025298856509228, "clip_ratio/low_min": 0.00024469397564341, "clip_ratio/region_mean": 0.0020841485675191507, "completions/clipped_ratio": 0.0, "completions/max_length": 639.4, "completions/max_terminated_length": 639.4, "completions/mean_length": 212.1471405029297, "completions/mean_terminated_length": 212.1471405029297, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.03447681564527253, "epoch": 0.8815426997245179, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 147608885.0, "reward": 8.08702130317688, "reward_std": 0.4030572921037674, "rewards/combined_reward/mean": 8.0870210647583, "rewards/combined_reward/std": 2.2266155242919923, "step": 1440 }, { "clip_ratio/high_max": 0.0018324234437992952, "clip_ratio/high_mean": 0.0011463960453208224, "clip_ratio/low_mean": 0.0008555376519022199, "clip_ratio/low_min": 0.00027054825137990215, "clip_ratio/region_mean": 0.002001933695525319, "completions/clipped_ratio": 0.0, "completions/max_length": 656.4, "completions/max_terminated_length": 656.4, "completions/mean_length": 217.47656860351563, "completions/mean_terminated_length": 217.47656860351563, "completions/min_length": 49.1, "completions/min_terminated_length": 49.1, "entropy": 0.031104416043187182, "epoch": 0.8999081726354453, "frac_reward_zero_std": 0.14583333693444728, "grad_norm": 0.0101318359375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 150744155.0, "reward": 8.09107985496521, "reward_std": 0.3631019815802574, "rewards/combined_reward/mean": 8.091079187393188, "rewards/combined_reward/std": 2.112835156917572, "step": 1470 }, { "clip_ratio/high_max": 0.0023938165104482323, "clip_ratio/high_mean": 0.0014270928280893714, "clip_ratio/low_mean": 0.0010784296340716538, "clip_ratio/low_min": 0.0003061608578718733, "clip_ratio/region_mean": 0.0025055224522172164, "completions/clipped_ratio": 0.0, "completions/max_length": 616.1, "completions/max_terminated_length": 616.1, "completions/mean_length": 189.75521240234374, "completions/mean_terminated_length": 189.75521240234374, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03524117438743512, "epoch": 0.9182736455463728, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 153772975.0, "reward": 8.024082374572753, "reward_std": 0.43461503088474274, "rewards/combined_reward/mean": 8.024081945419312, "rewards/combined_reward/std": 2.1382607936859133, "step": 1500 }, { "clip_ratio/high_max": 0.0020513815310550854, "clip_ratio/high_mean": 0.0011982314345611181, "clip_ratio/low_mean": 0.0007099811630420542, "clip_ratio/low_min": 0.00018517336114503752, "clip_ratio/region_mean": 0.001908212598937098, "completions/clipped_ratio": 0.0, "completions/max_length": 592.5, "completions/max_terminated_length": 592.5, "completions/mean_length": 205.79792022705078, "completions/mean_terminated_length": 205.79792022705078, "completions/min_length": 45.7, "completions/min_terminated_length": 45.7, "entropy": 0.02982318641928335, "epoch": 0.9366391184573003, "frac_reward_zero_std": 0.2125000052154064, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 156863399.0, "reward": 8.159046125411987, "reward_std": 0.28465825915336607, "rewards/combined_reward/mean": 8.159045648574828, "rewards/combined_reward/std": 2.3054054498672487, "step": 1530 }, { "clip_ratio/high_max": 0.001924379726309174, "clip_ratio/high_mean": 0.00114816524668034, "clip_ratio/low_mean": 0.0007961817152439229, "clip_ratio/low_min": 0.0002659123538857481, "clip_ratio/region_mean": 0.001944346949555135, "completions/clipped_ratio": 0.0, "completions/max_length": 604.1, "completions/max_terminated_length": 604.1, "completions/mean_length": 209.3807342529297, "completions/mean_terminated_length": 209.3807342529297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.031839094264432787, "epoch": 0.9550045913682277, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 159967581.0, "reward": 7.814377164840698, "reward_std": 0.36381800174713136, "rewards/combined_reward/mean": 7.814376783370972, "rewards/combined_reward/std": 2.4122848987579344, "step": 1560 }, { "clip_ratio/high_max": 0.0021151445369469004, "clip_ratio/high_mean": 0.001160936821785678, "clip_ratio/low_mean": 0.000825352026974239, "clip_ratio/low_min": 0.00021742046907699357, "clip_ratio/region_mean": 0.001986288835784459, "completions/clipped_ratio": 0.0, "completions/max_length": 617.4, "completions/max_terminated_length": 617.4, "completions/mean_length": 194.60000305175782, "completions/mean_terminated_length": 194.60000305175782, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.030347547059257827, "epoch": 0.9733700642791552, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 163015005.0, "reward": 8.083687686920166, "reward_std": 0.29379695653915405, "rewards/combined_reward/mean": 8.083687591552735, "rewards/combined_reward/std": 2.300224041938782, "step": 1590 }, { "clip_ratio/high_max": 0.001847240347706247, "clip_ratio/high_mean": 0.001069208953413181, "clip_ratio/low_mean": 0.0009093076396917847, "clip_ratio/low_min": 0.00029922581549423434, "clip_ratio/region_mean": 0.001978516616024232, "completions/clipped_ratio": 0.0, "completions/max_length": 615.4, "completions/max_terminated_length": 615.4, "completions/mean_length": 201.84036865234376, "completions/mean_terminated_length": 201.84036865234376, "completions/min_length": 45.5, "completions/min_terminated_length": 45.5, "entropy": 0.028152926017840703, "epoch": 0.9917355371900827, "frac_reward_zero_std": 0.1583333346992731, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 166090232.0, "reward": 7.961093854904175, "reward_std": 0.3353623315691948, "rewards/combined_reward/mean": 7.9610936641693115, "rewards/combined_reward/std": 2.3578851580619813, "step": 1620 }, { "clip_ratio/high_max": 0.0026378453912911937, "clip_ratio/high_mean": 0.0014017448585206184, "clip_ratio/low_mean": 0.0007972488769525426, "clip_ratio/low_min": 0.0002500811861571461, "clip_ratio/region_mean": 0.0021989937310335937, "completions/clipped_ratio": 0.0, "completions/max_length": 522.3, "completions/max_terminated_length": 522.3, "completions/mean_length": 205.4367248535156, "completions/mean_terminated_length": 205.4367248535156, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.031232609145217024, "epoch": 1.009794918885828, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 169179269.0, "reward": 8.17302508354187, "reward_std": 0.3992866724729538, "rewards/combined_reward/mean": 8.17302484512329, "rewards/combined_reward/std": 2.1745608866214754, "step": 1650 }, { "clip_ratio/high_max": 0.0023084808606654407, "clip_ratio/high_mean": 0.0014584960964081498, "clip_ratio/low_mean": 0.0013175203814171255, "clip_ratio/low_min": 0.00047267080517485736, "clip_ratio/region_mean": 0.002776016484616169, "completions/clipped_ratio": 0.0, "completions/max_length": 495.3, "completions/max_terminated_length": 495.3, "completions/mean_length": 191.54271392822267, "completions/mean_terminated_length": 191.54271392822267, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.031193043493355315, "epoch": 1.0281603917967554, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 172214953.0, "reward": 7.978023052215576, "reward_std": 0.3222086489200592, "rewards/combined_reward/mean": 7.978022909164428, "rewards/combined_reward/std": 2.310355567932129, "step": 1680 }, { "clip_ratio/high_max": 0.0028589733997553897, "clip_ratio/high_mean": 0.0016839676534194344, "clip_ratio/low_mean": 0.0012348661959549645, "clip_ratio/low_min": 0.0003763720960705541, "clip_ratio/region_mean": 0.0029188338274252603, "completions/clipped_ratio": 0.0, "completions/max_length": 579.2, "completions/max_terminated_length": 579.2, "completions/mean_length": 211.9401092529297, "completions/mean_terminated_length": 211.9401092529297, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.036311814375221727, "epoch": 1.046525864707683, "frac_reward_zero_std": 0.15000000596046448, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 175328963.0, "reward": 7.999973249435425, "reward_std": 0.3870839595794678, "rewards/combined_reward/mean": 7.99997296333313, "rewards/combined_reward/std": 2.2451637864112852, "step": 1710 }, { "clip_ratio/high_max": 0.002243271764018573, "clip_ratio/high_mean": 0.0013352544541703538, "clip_ratio/low_mean": 0.0010129570757271722, "clip_ratio/low_min": 0.0002704890214469439, "clip_ratio/region_mean": 0.002348211531837781, "completions/clipped_ratio": 0.0, "completions/max_length": 488.7, "completions/max_terminated_length": 488.7, "completions/mean_length": 203.3664123535156, "completions/mean_terminated_length": 203.3664123535156, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.027273702435195446, "epoch": 1.0648913376186104, "frac_reward_zero_std": 0.13750000447034835, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 178410050.0, "reward": 8.234262466430664, "reward_std": 0.3307895094156265, "rewards/combined_reward/mean": 8.234261846542358, "rewards/combined_reward/std": 2.1582916498184206, "step": 1740 }, { "clip_ratio/high_max": 0.0025796565372729675, "clip_ratio/high_mean": 0.0013540853087761207, "clip_ratio/low_mean": 0.0010891686063284093, "clip_ratio/low_min": 0.00026517792624266195, "clip_ratio/region_mean": 0.0024432539095869286, "completions/clipped_ratio": 0.0, "completions/max_length": 548.2, "completions/max_terminated_length": 548.2, "completions/mean_length": 195.88906860351562, "completions/mean_terminated_length": 195.88906860351562, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030365773709490895, "epoch": 1.0832568105295377, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.006988525390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 181462424.0, "reward": 8.225980854034423, "reward_std": 0.33549507707357407, "rewards/combined_reward/mean": 8.225980615615844, "rewards/combined_reward/std": 1.9718674063682555, "step": 1770 }, { "clip_ratio/high_max": 0.0026900336107549567, "clip_ratio/high_mean": 0.0015414831575374895, "clip_ratio/low_mean": 0.001024144281958191, "clip_ratio/low_min": 0.0003565661820175592, "clip_ratio/region_mean": 0.00256562743500884, "completions/clipped_ratio": 0.0, "completions/max_length": 740.3, "completions/max_terminated_length": 740.3, "completions/mean_length": 207.61172332763672, "completions/mean_terminated_length": 207.61172332763672, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.034605284392212825, "epoch": 1.1016222834404652, "frac_reward_zero_std": 0.13750000149011612, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 184559813.0, "reward": 7.892181396484375, "reward_std": 0.3990317821502686, "rewards/combined_reward/mean": 7.892181062698365, "rewards/combined_reward/std": 2.2660832524299623, "step": 1800 }, { "clip_ratio/high_max": 0.00254516903514741, "clip_ratio/high_mean": 0.0014866135614283848, "clip_ratio/low_mean": 0.0009874311481932333, "clip_ratio/low_min": 0.0003794685149235496, "clip_ratio/region_mean": 0.0024740447183527674, "completions/clipped_ratio": 0.0, "completions/max_length": 677.9, "completions/max_terminated_length": 677.9, "completions/mean_length": 215.1088607788086, "completions/mean_terminated_length": 215.1088607788086, "completions/min_length": 48.6, "completions/min_terminated_length": 48.6, "entropy": 0.03513814099133015, "epoch": 1.1199877563513927, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 187685991.0, "reward": 7.862548446655273, "reward_std": 0.36118823289871216, "rewards/combined_reward/mean": 7.862548303604126, "rewards/combined_reward/std": 2.6146493673324587, "step": 1830 }, { "clip_ratio/high_max": 0.0024625309199715656, "clip_ratio/high_mean": 0.0013821574562219515, "clip_ratio/low_mean": 0.0010081118271045852, "clip_ratio/low_min": 0.0002770640584155141, "clip_ratio/region_mean": 0.002390269273989058, "completions/clipped_ratio": 0.0, "completions/max_length": 619.1, "completions/max_terminated_length": 619.1, "completions/mean_length": 189.39974517822264, "completions/mean_terminated_length": 189.39974517822264, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0314952748361975, "epoch": 1.1383532292623202, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.00579833984375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 190713446.0, "reward": 7.961237907409668, "reward_std": 0.2653660446405411, "rewards/combined_reward/mean": 7.961237668991089, "rewards/combined_reward/std": 2.3187624335289003, "step": 1860 }, { "clip_ratio/high_max": 0.002394012764367896, "clip_ratio/high_mean": 0.0012987459883637106, "clip_ratio/low_mean": 0.0010693162638441814, "clip_ratio/low_min": 0.0003574072805349715, "clip_ratio/region_mean": 0.0023680622589987857, "completions/clipped_ratio": 0.0, "completions/max_length": 618.5, "completions/max_terminated_length": 618.5, "completions/mean_length": 202.87187957763672, "completions/mean_terminated_length": 202.87187957763672, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.030985799074793857, "epoch": 1.1567187021732477, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 193792634.0, "reward": 8.041134071350097, "reward_std": 0.4226482570171356, "rewards/combined_reward/mean": 8.041133689880372, "rewards/combined_reward/std": 2.186667114496231, "step": 1890 }, { "clip_ratio/high_max": 0.002431436919141561, "clip_ratio/high_mean": 0.0014043091953984306, "clip_ratio/low_mean": 0.0008318756178293066, "clip_ratio/low_min": 0.0002537611374767342, "clip_ratio/region_mean": 0.002236184807406971, "completions/clipped_ratio": 0.0, "completions/max_length": 526.8, "completions/max_terminated_length": 526.8, "completions/mean_length": 206.40365142822264, "completions/mean_terminated_length": 206.40365142822264, "completions/min_length": 46.2, "completions/min_terminated_length": 46.2, "entropy": 0.02784327861542503, "epoch": 1.1750841750841752, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.024169921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 196885384.0, "reward": 8.051492309570312, "reward_std": 0.34346550703048706, "rewards/combined_reward/mean": 8.051492023468018, "rewards/combined_reward/std": 2.2653783440589903, "step": 1920 }, { "clip_ratio/high_max": 0.002355142892823399, "clip_ratio/high_mean": 0.0014093171583226649, "clip_ratio/low_mean": 0.0009466349748739352, "clip_ratio/low_min": 0.00023793850389968913, "clip_ratio/region_mean": 0.0023559521388961, "completions/clipped_ratio": 0.0, "completions/max_length": 522.4, "completions/max_terminated_length": 522.4, "completions/mean_length": 199.36536865234376, "completions/mean_terminated_length": 199.36536865234376, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03002636181966712, "epoch": 1.1934496479951024, "frac_reward_zero_std": 0.18333333767950535, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 199951107.0, "reward": 7.98937954902649, "reward_std": 0.2641743287444115, "rewards/combined_reward/mean": 7.989379262924194, "rewards/combined_reward/std": 2.3922903060913088, "step": 1950 }, { "clip_ratio/high_max": 0.0023929632962487327, "clip_ratio/high_mean": 0.0013180818787077442, "clip_ratio/low_mean": 0.0009481080422119703, "clip_ratio/low_min": 0.00022901383805825997, "clip_ratio/region_mean": 0.002266189910490842, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 195.5846405029297, "completions/mean_terminated_length": 195.5846405029297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028338880681743225, "epoch": 1.21181512090603, "frac_reward_zero_std": 0.18333333507180213, "grad_norm": 0.005279541015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 203002312.0, "reward": 8.084173250198365, "reward_std": 0.3409020766615868, "rewards/combined_reward/mean": 8.084173154830932, "rewards/combined_reward/std": 2.3569565773010255, "step": 1980 }, { "clip_ratio/high_max": 0.002849062965348518, "clip_ratio/high_mean": 0.001635522501116308, "clip_ratio/low_mean": 0.0010733801702372148, "clip_ratio/low_min": 0.00036468363299112146, "clip_ratio/region_mean": 0.0027089026659571874, "completions/clipped_ratio": 0.0, "completions/max_length": 619.3, "completions/max_terminated_length": 619.3, "completions/mean_length": 202.01614990234376, "completions/mean_terminated_length": 202.01614990234376, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.03378065534246465, "epoch": 1.2301805938169574, "frac_reward_zero_std": 0.15416666828095912, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 206078214.0, "reward": 8.049263334274292, "reward_std": 0.38385118544101715, "rewards/combined_reward/mean": 8.049263143539429, "rewards/combined_reward/std": 2.235533571243286, "step": 2010 }, { "clip_ratio/high_max": 0.002582704926317092, "clip_ratio/high_mean": 0.0014030443922820267, "clip_ratio/low_mean": 0.001040563203059719, "clip_ratio/low_min": 0.0002656340984685812, "clip_ratio/region_mean": 0.002443607580304767, "completions/clipped_ratio": 0.0, "completions/max_length": 615.5, "completions/max_terminated_length": 615.5, "completions/mean_length": 196.29427490234374, "completions/mean_terminated_length": 196.29427490234374, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.029009897292902074, "epoch": 1.248546066727885, "frac_reward_zero_std": 0.15000000298023225, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 209132144.0, "reward": 8.07990870475769, "reward_std": 0.3697445645928383, "rewards/combined_reward/mean": 8.079908418655396, "rewards/combined_reward/std": 2.1825843930244444, "step": 2040 }, { "clip_ratio/high_max": 0.002421984917600639, "clip_ratio/high_mean": 0.0013921495222651477, "clip_ratio/low_mean": 0.0009136913968783726, "clip_ratio/low_min": 0.00023133103798803252, "clip_ratio/region_mean": 0.002305840932482776, "completions/clipped_ratio": 0.0, "completions/max_length": 573.4, "completions/max_terminated_length": 573.4, "completions/mean_length": 197.62240142822264, "completions/mean_terminated_length": 197.62240142822264, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.029476982774212956, "epoch": 1.2669115396388124, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.005218505859375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 212191174.0, "reward": 8.082415962219239, "reward_std": 0.29564848393201826, "rewards/combined_reward/mean": 8.082415771484374, "rewards/combined_reward/std": 2.234119880199432, "step": 2070 }, { "clip_ratio/high_max": 0.0026495455119099156, "clip_ratio/high_mean": 0.0014595542930085988, "clip_ratio/low_mean": 0.0009755512626725249, "clip_ratio/low_min": 0.00017299341537485208, "clip_ratio/region_mean": 0.0024351055690203793, "completions/clipped_ratio": 0.0, "completions/max_length": 657.1, "completions/max_terminated_length": 657.1, "completions/mean_length": 189.39010772705078, "completions/mean_terminated_length": 189.39010772705078, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03329295050352812, "epoch": 1.2852770125497397, "frac_reward_zero_std": 0.20416667349636555, "grad_norm": 0.025390625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 215218592.0, "reward": 7.946870279312134, "reward_std": 0.3564735472202301, "rewards/combined_reward/mean": 7.9468700885772705, "rewards/combined_reward/std": 2.3863738894462587, "step": 2100 }, { "clip_ratio/high_max": 0.0028313372255070134, "clip_ratio/high_mean": 0.001543560571250661, "clip_ratio/low_mean": 0.0010128551270099707, "clip_ratio/low_min": 0.000293470832790869, "clip_ratio/region_mean": 0.002556415705475956, "completions/clipped_ratio": 0.0, "completions/max_length": 609.7, "completions/max_terminated_length": 609.7, "completions/mean_length": 204.25443115234376, "completions/mean_terminated_length": 204.25443115234376, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02987209097482264, "epoch": 1.3036424854606672, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0235595703125, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 218303089.0, "reward": 7.873810911178589, "reward_std": 0.49495945274829867, "rewards/combined_reward/mean": 7.873810720443726, "rewards/combined_reward/std": 2.349976146221161, "step": 2130 }, { "clip_ratio/high_max": 0.0027523085116020712, "clip_ratio/high_mean": 0.0014988247033519049, "clip_ratio/low_mean": 0.0009871584678573222, "clip_ratio/low_min": 0.00029718875860756575, "clip_ratio/region_mean": 0.002485983166843653, "completions/clipped_ratio": 0.0, "completions/max_length": 611.9, "completions/max_terminated_length": 611.9, "completions/mean_length": 205.22760925292968, "completions/mean_terminated_length": 205.22760925292968, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.031543298003574215, "epoch": 1.3220079583715947, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 221391323.0, "reward": 8.21387529373169, "reward_std": 0.4003839701414108, "rewards/combined_reward/mean": 8.21387505531311, "rewards/combined_reward/std": 2.0670359313488005, "step": 2160 }, { "clip_ratio/high_max": 0.002586226168690094, "clip_ratio/high_mean": 0.0014906070340657608, "clip_ratio/low_mean": 0.0011603268125327305, "clip_ratio/low_min": 0.00034156959057630353, "clip_ratio/region_mean": 0.0026509338727919384, "completions/clipped_ratio": 0.0, "completions/max_length": 559.8, "completions/max_terminated_length": 559.8, "completions/mean_length": 206.08568267822267, "completions/mean_terminated_length": 206.08568267822267, "completions/min_length": 46.5, "completions/min_terminated_length": 46.5, "entropy": 0.03595255493807296, "epoch": 1.3403734312825222, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 224482852.0, "reward": 8.000640106201171, "reward_std": 0.36298624277114866, "rewards/combined_reward/mean": 8.000639915466309, "rewards/combined_reward/std": 2.5115744829177857, "step": 2190 }, { "clip_ratio/high_max": 0.002347306678711902, "clip_ratio/high_mean": 0.0013918503398599568, "clip_ratio/low_mean": 0.0009283636261898209, "clip_ratio/low_min": 0.0002616868099721614, "clip_ratio/region_mean": 0.0023202139598045807, "completions/clipped_ratio": 0.0, "completions/max_length": 620.7, "completions/max_terminated_length": 620.7, "completions/mean_length": 204.56693420410156, "completions/mean_terminated_length": 204.56693420410156, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.031378321411708994, "epoch": 1.3587389041934497, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 227568549.0, "reward": 8.194380807876588, "reward_std": 0.3646244630217552, "rewards/combined_reward/mean": 8.194380378723144, "rewards/combined_reward/std": 2.1398088693618775, "step": 2220 }, { "clip_ratio/high_max": 0.002533982332776456, "clip_ratio/high_mean": 0.001468075599890047, "clip_ratio/low_mean": 0.0009899798745512574, "clip_ratio/low_min": 0.0002981170568091329, "clip_ratio/region_mean": 0.0024580554648612938, "completions/clipped_ratio": 0.0, "completions/max_length": 592.3, "completions/max_terminated_length": 592.3, "completions/mean_length": 203.04662017822267, "completions/mean_terminated_length": 203.04662017822267, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03200453007593751, "epoch": 1.3771043771043772, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 230648408.0, "reward": 7.9668128967285154, "reward_std": 0.3462532043457031, "rewards/combined_reward/mean": 7.966812610626221, "rewards/combined_reward/std": 2.371821975708008, "step": 2250 }, { "clip_ratio/high_max": 0.002723057131030752, "clip_ratio/high_mean": 0.0016616024520772043, "clip_ratio/low_mean": 0.0013893289481832956, "clip_ratio/low_min": 0.00042196267095278016, "clip_ratio/region_mean": 0.0030509313840108615, "completions/clipped_ratio": 0.0, "completions/max_length": 566.3, "completions/max_terminated_length": 566.3, "completions/mean_length": 207.7205795288086, "completions/mean_terminated_length": 207.7205795288086, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.0363157723719875, "epoch": 1.3954698500153047, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 233746215.0, "reward": 7.639039421081543, "reward_std": 0.42567554861307144, "rewards/combined_reward/mean": 7.639039182662964, "rewards/combined_reward/std": 2.5516008853912355, "step": 2280 }, { "clip_ratio/high_max": 0.0027178450114054916, "clip_ratio/high_mean": 0.001420725736291691, "clip_ratio/low_mean": 0.0009931883843819378, "clip_ratio/low_min": 0.00025444680383467734, "clip_ratio/region_mean": 0.002413914126130597, "completions/clipped_ratio": 0.0, "completions/max_length": 537.1, "completions/max_terminated_length": 537.1, "completions/mean_length": 195.8815155029297, "completions/mean_terminated_length": 195.8815155029297, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.029116225723798076, "epoch": 1.4138353229262322, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 236798560.0, "reward": 7.902426767349243, "reward_std": 0.3571103274822235, "rewards/combined_reward/mean": 7.9024265766143795, "rewards/combined_reward/std": 2.325634753704071, "step": 2310 }, { "clip_ratio/high_max": 0.0019280185675597749, "clip_ratio/high_mean": 0.001165979286027626, "clip_ratio/low_mean": 0.0009077533976475631, "clip_ratio/low_min": 0.0002470596162311267, "clip_ratio/region_mean": 0.0020737326859186092, "completions/clipped_ratio": 0.0, "completions/max_length": 689.8, "completions/max_terminated_length": 689.8, "completions/mean_length": 207.94870300292968, "completions/mean_terminated_length": 207.94870300292968, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.02850599765467147, "epoch": 1.4322007958371594, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 239897243.0, "reward": 7.842256212234497, "reward_std": 0.3329267457127571, "rewards/combined_reward/mean": 7.842255783081055, "rewards/combined_reward/std": 2.655168962478638, "step": 2340 }, { "clip_ratio/high_max": 0.002345844705996569, "clip_ratio/high_mean": 0.0013676094364200252, "clip_ratio/low_mean": 0.0010573526046452268, "clip_ratio/low_min": 0.0002996241637447383, "clip_ratio/region_mean": 0.002424962014871805, "completions/clipped_ratio": 0.0, "completions/max_length": 668.9, "completions/max_terminated_length": 668.9, "completions/mean_length": 206.07240142822266, "completions/mean_terminated_length": 206.07240142822266, "completions/min_length": 46.9, "completions/min_terminated_length": 46.9, "entropy": 0.033145268773660065, "epoch": 1.450566268748087, "frac_reward_zero_std": 0.18750000707805156, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 242988721.0, "reward": 8.17882742881775, "reward_std": 0.3625470131635666, "rewards/combined_reward/mean": 8.178827142715454, "rewards/combined_reward/std": 2.3492067098617553, "step": 2370 }, { "clip_ratio/high_max": 0.0029163743660319597, "clip_ratio/high_mean": 0.001694265996775357, "clip_ratio/low_mean": 0.001122715271412744, "clip_ratio/low_min": 0.00026123497567217174, "clip_ratio/region_mean": 0.0028169812775255803, "completions/clipped_ratio": 0.0, "completions/max_length": 529.5, "completions/max_terminated_length": 529.5, "completions/mean_length": 198.13750457763672, "completions/mean_terminated_length": 198.13750457763672, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03350481181405485, "epoch": 1.4689317416590144, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 246049729.0, "reward": 7.92017970085144, "reward_std": 0.3565027922391891, "rewards/combined_reward/mean": 7.920179414749145, "rewards/combined_reward/std": 2.2732934951782227, "step": 2400 }, { "clip_ratio/high_max": 0.001715045036689844, "clip_ratio/high_mean": 0.0009998808901097314, "clip_ratio/low_mean": 0.0008196418659887664, "clip_ratio/low_min": 0.00024973992840386926, "clip_ratio/region_mean": 0.0018195227598577428, "completions/clipped_ratio": 0.0, "completions/max_length": 565.8, "completions/max_terminated_length": 565.8, "completions/mean_length": 204.03880767822267, "completions/mean_terminated_length": 204.03880767822267, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.030072947824373843, "epoch": 1.487297214569942, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 249133398.0, "reward": 7.92440881729126, "reward_std": 0.418508405983448, "rewards/combined_reward/mean": 7.924408340454102, "rewards/combined_reward/std": 2.214128279685974, "step": 2430 }, { "clip_ratio/high_max": 0.0019889042790358267, "clip_ratio/high_mean": 0.0011504420102331399, "clip_ratio/low_mean": 0.0007639368074402834, "clip_ratio/low_min": 0.00019526908339078847, "clip_ratio/region_mean": 0.001914378812458987, "completions/clipped_ratio": 0.0, "completions/max_length": 586.3, "completions/max_terminated_length": 586.3, "completions/mean_length": 203.9416732788086, "completions/mean_terminated_length": 203.9416732788086, "completions/min_length": 48.4, "completions/min_terminated_length": 48.4, "entropy": 0.030919911774496236, "epoch": 1.5056626874808692, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 252216694.0, "reward": 7.80237045288086, "reward_std": 0.39187341928482056, "rewards/combined_reward/mean": 7.80237045288086, "rewards/combined_reward/std": 2.4975873947143556, "step": 2460 }, { "clip_ratio/high_max": 0.0019456488732733609, "clip_ratio/high_mean": 0.0011542707798071206, "clip_ratio/low_mean": 0.0008219932990565818, "clip_ratio/low_min": 0.00025135041020500165, "clip_ratio/region_mean": 0.001976264076074585, "completions/clipped_ratio": 0.0, "completions/max_length": 611.5, "completions/max_terminated_length": 611.5, "completions/mean_length": 202.28906707763673, "completions/mean_terminated_length": 202.28906707763673, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.029948565643280745, "epoch": 1.5240281603917967, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 255293644.0, "reward": 7.835418796539306, "reward_std": 0.32145274728536605, "rewards/combined_reward/mean": 7.835418748855591, "rewards/combined_reward/std": 2.4204742312431335, "step": 2490 }, { "clip_ratio/high_max": 0.0020958329337493827, "clip_ratio/high_mean": 0.0012271287857098894, "clip_ratio/low_mean": 0.0008954514073896765, "clip_ratio/low_min": 0.00026732577510604945, "clip_ratio/region_mean": 0.002122580165935991, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 210.80964202880858, "completions/mean_terminated_length": 210.80964202880858, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.03271008868080874, "epoch": 1.5423936333027242, "frac_reward_zero_std": 0.1583333395421505, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 258403313.0, "reward": 7.760156059265137, "reward_std": 0.4169061690568924, "rewards/combined_reward/mean": 7.760155916213989, "rewards/combined_reward/std": 2.5645476460456846, "step": 2520 }, { "clip_ratio/high_max": 0.0021703777010164535, "clip_ratio/high_mean": 0.0012445400386544255, "clip_ratio/low_mean": 0.000831753575281861, "clip_ratio/low_min": 0.00018968698980946405, "clip_ratio/region_mean": 0.002076293590168158, "completions/clipped_ratio": 0.0, "completions/max_length": 611.7, "completions/max_terminated_length": 611.7, "completions/mean_length": 197.40286712646486, "completions/mean_terminated_length": 197.40286712646486, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03004558289734026, "epoch": 1.5607591062136517, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0032, "num_tokens": 261461500.0, "reward": 8.038401460647583, "reward_std": 0.476505346596241, "rewards/combined_reward/mean": 8.038401126861572, "rewards/combined_reward/std": 2.2272711396217346, "step": 2550 }, { "clip_ratio/high_max": 0.0018087226160181066, "clip_ratio/high_mean": 0.001009383638544629, "clip_ratio/low_mean": 0.0008766352189316724, "clip_ratio/low_min": 0.00027314005856169386, "clip_ratio/region_mean": 0.0018860188584464291, "completions/clipped_ratio": 0.0, "completions/max_length": 500.7, "completions/max_terminated_length": 500.7, "completions/mean_length": 195.19766082763672, "completions/mean_terminated_length": 195.19766082763672, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.029871799486378827, "epoch": 1.5791245791245792, "frac_reward_zero_std": 0.15416667237877846, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 264511219.0, "reward": 8.038916301727294, "reward_std": 0.3228034809231758, "rewards/combined_reward/mean": 8.038915729522705, "rewards/combined_reward/std": 2.269067847728729, "step": 2580 }, { "clip_ratio/high_max": 0.0022288179010502064, "clip_ratio/high_mean": 0.001268816123168411, "clip_ratio/low_mean": 0.0008975474728989259, "clip_ratio/low_min": 0.0002794755535433069, "clip_ratio/region_mean": 0.0021663635833344113, "completions/clipped_ratio": 0.0, "completions/max_length": 620.4, "completions/max_terminated_length": 620.4, "completions/mean_length": 191.12891235351563, "completions/mean_terminated_length": 191.12891235351563, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.03397136890950302, "epoch": 1.5974900520355066, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 267545314.0, "reward": 8.063913202285766, "reward_std": 0.38206731230020524, "rewards/combined_reward/mean": 8.063912916183472, "rewards/combined_reward/std": 2.2844601035118104, "step": 2610 }, { "clip_ratio/high_max": 0.0019982606662476126, "clip_ratio/high_mean": 0.0012319115875773909, "clip_ratio/low_mean": 0.0008621523914674375, "clip_ratio/low_min": 0.0002622997337311972, "clip_ratio/region_mean": 0.0020940639680096258, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 202.4924514770508, "completions/mean_terminated_length": 202.4924514770508, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.030762230053854485, "epoch": 1.6158555249464341, "frac_reward_zero_std": 0.19166667573153973, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 270623045.0, "reward": 8.11420316696167, "reward_std": 0.3650542140007019, "rewards/combined_reward/mean": 8.11420259475708, "rewards/combined_reward/std": 2.2436538219451903, "step": 2640 }, { "clip_ratio/high_max": 0.0018537834150871883, "clip_ratio/high_mean": 0.0011615452634335572, "clip_ratio/low_mean": 0.0009258029560442083, "clip_ratio/low_min": 0.00027814039907146556, "clip_ratio/region_mean": 0.002087348224692202, "completions/clipped_ratio": 0.0, "completions/max_length": 696.6, "completions/max_terminated_length": 696.6, "completions/mean_length": 191.68724517822267, "completions/mean_terminated_length": 191.68724517822267, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03179402017655472, "epoch": 1.6342209978573616, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 273659284.0, "reward": 7.900281000137329, "reward_std": 0.4007427215576172, "rewards/combined_reward/mean": 7.90028076171875, "rewards/combined_reward/std": 2.531586754322052, "step": 2670 }, { "clip_ratio/high_max": 0.0017818992518490026, "clip_ratio/high_mean": 0.0010655213587597244, "clip_ratio/low_mean": 0.0006358582640435391, "clip_ratio/low_min": 0.00019485876691760495, "clip_ratio/region_mean": 0.0017013796242584553, "completions/clipped_ratio": 0.0, "completions/max_length": 524.5, "completions/max_terminated_length": 524.5, "completions/mean_length": 188.74739990234374, "completions/mean_terminated_length": 188.74739990234374, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.028859341928424934, "epoch": 1.652586470768289, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 276684234.0, "reward": 7.948795223236084, "reward_std": 0.4388988882303238, "rewards/combined_reward/mean": 7.948795032501221, "rewards/combined_reward/std": 2.26337993144989, "step": 2700 }, { "clip_ratio/high_max": 0.0022360322897535903, "clip_ratio/high_mean": 0.0011552742324662782, "clip_ratio/low_mean": 0.0009495423510088585, "clip_ratio/low_min": 0.000272480478209521, "clip_ratio/region_mean": 0.0021048165872343816, "completions/clipped_ratio": 0.0, "completions/max_length": 617.8, "completions/max_terminated_length": 617.8, "completions/mean_length": 206.0143280029297, "completions/mean_terminated_length": 206.0143280029297, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.030598011643936236, "epoch": 1.6709519436792164, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.003204345703125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 279775489.0, "reward": 7.868149614334106, "reward_std": 0.36714731231331826, "rewards/combined_reward/mean": 7.868149375915527, "rewards/combined_reward/std": 2.4306089997291567, "step": 2730 }, { "clip_ratio/high_max": 0.0019022035034140573, "clip_ratio/high_mean": 0.0011078421055572107, "clip_ratio/low_mean": 0.0006738288718527959, "clip_ratio/low_min": 0.0001755885229310176, "clip_ratio/region_mean": 0.0017816709878388792, "completions/clipped_ratio": 0.0, "completions/max_length": 510.4, "completions/max_terminated_length": 510.4, "completions/mean_length": 188.8557327270508, "completions/mean_terminated_length": 188.8557327270508, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.028884920741741857, "epoch": 1.689317416590144, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 282800855.0, "reward": 7.593391847610474, "reward_std": 0.3909902811050415, "rewards/combined_reward/mean": 7.5933915138244625, "rewards/combined_reward/std": 2.4906132221221924, "step": 2760 }, { "clip_ratio/high_max": 0.0021964017939656816, "clip_ratio/high_mean": 0.0012622291719405136, "clip_ratio/low_mean": 0.0008550863655424715, "clip_ratio/low_min": 0.0003408503873894612, "clip_ratio/region_mean": 0.0021173155158370114, "completions/clipped_ratio": 0.0, "completions/max_length": 642.2, "completions/max_terminated_length": 642.2, "completions/mean_length": 196.60677642822264, "completions/mean_terminated_length": 196.60677642822264, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02986331588278214, "epoch": 1.7076828895010712, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 285855985.0, "reward": 7.641462898254394, "reward_std": 0.40767830312252046, "rewards/combined_reward/mean": 7.641462898254394, "rewards/combined_reward/std": 2.6605490326881407, "step": 2790 }, { "clip_ratio/high_max": 0.0020044330585127076, "clip_ratio/high_mean": 0.00110363292672749, "clip_ratio/low_mean": 0.0008028148062294349, "clip_ratio/low_min": 0.00020776684298956145, "clip_ratio/region_mean": 0.0019064477300465418, "completions/clipped_ratio": 0.0, "completions/max_length": 497.9, "completions/max_terminated_length": 497.9, "completions/mean_length": 198.72552795410155, "completions/mean_terminated_length": 198.72552795410155, "completions/min_length": 47.5, "completions/min_terminated_length": 47.5, "entropy": 0.030650685634464027, "epoch": 1.7260483624119987, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 288919251.0, "reward": 8.076289653778076, "reward_std": 0.32201329320669175, "rewards/combined_reward/mean": 8.076289463043214, "rewards/combined_reward/std": 2.272538483142853, "step": 2820 }, { "clip_ratio/high_max": 0.002055777466739528, "clip_ratio/high_mean": 0.0011928419422474689, "clip_ratio/low_mean": 0.0007476838151584767, "clip_ratio/low_min": 0.00025830524682532994, "clip_ratio/region_mean": 0.0019405257582548075, "completions/clipped_ratio": 0.0, "completions/max_length": 619.9, "completions/max_terminated_length": 619.9, "completions/mean_length": 200.3947967529297, "completions/mean_terminated_length": 200.3947967529297, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.031035555611985424, "epoch": 1.7444138353229262, "frac_reward_zero_std": 0.15000000596046448, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 291988927.0, "reward": 8.02885036468506, "reward_std": 0.397235581278801, "rewards/combined_reward/mean": 8.028850317001343, "rewards/combined_reward/std": 2.387756717205048, "step": 2850 }, { "clip_ratio/high_max": 0.001939509065656845, "clip_ratio/high_mean": 0.0011942439775642318, "clip_ratio/low_mean": 0.0008201938236500913, "clip_ratio/low_min": 0.00025946315120866834, "clip_ratio/region_mean": 0.002014437831530813, "completions/clipped_ratio": 0.0, "completions/max_length": 622.5, "completions/max_terminated_length": 622.5, "completions/mean_length": 206.60521545410157, "completions/mean_terminated_length": 206.60521545410157, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.03139836806803942, "epoch": 1.7627793082338536, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.02587890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 295082451.0, "reward": 8.01548137664795, "reward_std": 0.3635264039039612, "rewards/combined_reward/mean": 8.015481185913085, "rewards/combined_reward/std": 2.3814555644989013, "step": 2880 }, { "clip_ratio/high_max": 0.0016605210689400944, "clip_ratio/high_mean": 0.0009404748006393977, "clip_ratio/low_mean": 0.0006891741961832546, "clip_ratio/low_min": 0.00021507224397889028, "clip_ratio/region_mean": 0.001629649010525706, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 194.13985137939454, "completions/mean_terminated_length": 194.13985137939454, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.024972609026978412, "epoch": 1.7811447811447811, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 298128108.0, "reward": 8.214980792999267, "reward_std": 0.4014619827270508, "rewards/combined_reward/mean": 8.214980220794677, "rewards/combined_reward/std": 2.0835484623908997, "step": 2910 }, { "clip_ratio/high_max": 0.0021403315074470203, "clip_ratio/high_mean": 0.0012256344707566313, "clip_ratio/low_mean": 0.0007606657380771746, "clip_ratio/low_min": 0.00018110804206420046, "clip_ratio/region_mean": 0.001986300202164178, "completions/clipped_ratio": 0.0, "completions/max_length": 632.6, "completions/max_terminated_length": 632.6, "completions/mean_length": 211.09818115234376, "completions/mean_terminated_length": 211.09818115234376, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.03239353181173404, "epoch": 1.7995102540557086, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 301238885.0, "reward": 8.048570346832275, "reward_std": 0.3699219077825546, "rewards/combined_reward/mean": 8.048570013046264, "rewards/combined_reward/std": 2.356692683696747, "step": 2940 }, { "clip_ratio/high_max": 0.0022189425285129498, "clip_ratio/high_mean": 0.0013083622373718147, "clip_ratio/low_mean": 0.0008259118498244788, "clip_ratio/low_min": 0.00023597468571097125, "clip_ratio/region_mean": 0.0021342740816180593, "completions/clipped_ratio": 0.0, "completions/max_length": 661.5, "completions/max_terminated_length": 661.5, "completions/mean_length": 208.44609832763672, "completions/mean_terminated_length": 208.44609832763672, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.03341625835746527, "epoch": 1.8178757269666361, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 304339478.0, "reward": 7.916926622390747, "reward_std": 0.4132001921534538, "rewards/combined_reward/mean": 7.916926431655884, "rewards/combined_reward/std": 2.2656158804893494, "step": 2970 }, { "clip_ratio/high_max": 0.002028170562698506, "clip_ratio/high_mean": 0.0012314242926853088, "clip_ratio/low_mean": 0.0007943610021053852, "clip_ratio/low_min": 0.0002534084948517072, "clip_ratio/region_mean": 0.0020257853083118487, "completions/clipped_ratio": 0.0, "completions/max_length": 601.2, "completions/max_terminated_length": 601.2, "completions/mean_length": 192.65208892822267, "completions/mean_terminated_length": 192.65208892822267, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03034144329528014, "epoch": 1.8362411998775636, "frac_reward_zero_std": 0.22083333879709244, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 307379422.0, "reward": 7.715467882156372, "reward_std": 0.399223130941391, "rewards/combined_reward/mean": 7.715467834472657, "rewards/combined_reward/std": 2.5660332918167112, "step": 3000 }, { "clip_ratio/high_max": 0.001616455056743386, "clip_ratio/high_mean": 0.0008959133473884624, "clip_ratio/low_mean": 0.0007587190180856851, "clip_ratio/low_min": 0.00017424074224739645, "clip_ratio/region_mean": 0.0016546323626243976, "completions/clipped_ratio": 0.0, "completions/max_length": 536.1, "completions/max_terminated_length": 536.1, "completions/mean_length": 193.88750610351562, "completions/mean_terminated_length": 193.88750610351562, "completions/min_length": 46.1, "completions/min_terminated_length": 46.1, "entropy": 0.026093621094090243, "epoch": 1.8546066727884911, "frac_reward_zero_std": 0.17083334028720856, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 310424110.0, "reward": 8.073449182510377, "reward_std": 0.33217342048883436, "rewards/combined_reward/mean": 8.073448848724365, "rewards/combined_reward/std": 2.3261872291564942, "step": 3030 }, { "clip_ratio/high_max": 0.0022083233731488385, "clip_ratio/high_mean": 0.0012707876936474348, "clip_ratio/low_mean": 0.0007065690054635828, "clip_ratio/low_min": 0.00015670985861409765, "clip_ratio/region_mean": 0.001977356710995082, "completions/clipped_ratio": 0.0, "completions/max_length": 585.3, "completions/max_terminated_length": 585.3, "completions/mean_length": 194.30964050292968, "completions/mean_terminated_length": 194.30964050292968, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03111668621810774, "epoch": 1.8729721456994184, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 313470419.0, "reward": 8.1059175491333, "reward_std": 0.3294979363679886, "rewards/combined_reward/mean": 8.105917406082153, "rewards/combined_reward/std": 2.2755979180335997, "step": 3060 }, { "clip_ratio/high_max": 0.001937716651688485, "clip_ratio/high_mean": 0.0011374628622434102, "clip_ratio/low_mean": 0.0008395296846477625, "clip_ratio/low_min": 0.0003360025019598349, "clip_ratio/region_mean": 0.001976992551741811, "completions/clipped_ratio": 0.0, "completions/max_length": 621.1, "completions/max_terminated_length": 621.1, "completions/mean_length": 217.01797332763672, "completions/mean_terminated_length": 217.01797332763672, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.034867001495634514, "epoch": 1.8913376186103459, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.00726318359375, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 316603928.0, "reward": 7.9157627582550045, "reward_std": 0.41175094842910764, "rewards/combined_reward/mean": 7.915762519836425, "rewards/combined_reward/std": 2.452712655067444, "step": 3090 }, { "clip_ratio/high_max": 0.0018715933047739478, "clip_ratio/high_mean": 0.0010689235023164656, "clip_ratio/low_mean": 0.0009667285920537931, "clip_ratio/low_min": 0.00025829696993848, "clip_ratio/region_mean": 0.002035652110983695, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 199.0158920288086, "completions/mean_terminated_length": 199.0158920288086, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.030959866723666588, "epoch": 1.9097030915212734, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.007171630859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 319668309.0, "reward": 8.071994733810424, "reward_std": 0.36481887102127075, "rewards/combined_reward/mean": 8.071994638442993, "rewards/combined_reward/std": 2.1674668073654173, "step": 3120 }, { "clip_ratio/high_max": 0.0020784744003321974, "clip_ratio/high_mean": 0.0013034821114464043, "clip_ratio/low_mean": 0.0008271059225080535, "clip_ratio/low_min": 0.00023410471840179525, "clip_ratio/region_mean": 0.002130588029588883, "completions/clipped_ratio": 0.0, "completions/max_length": 605.7, "completions/max_terminated_length": 605.7, "completions/mean_length": 206.12084045410157, "completions/mean_terminated_length": 206.12084045410157, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.03179230134313305, "epoch": 1.9280685644322006, "frac_reward_zero_std": 0.11250000186264515, "grad_norm": 0.008056640625, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 322759973.0, "reward": 7.747814178466797, "reward_std": 0.3783121943473816, "rewards/combined_reward/mean": 7.747813892364502, "rewards/combined_reward/std": 2.433961868286133, "step": 3150 }, { "clip_ratio/high_max": 0.0017183285073163765, "clip_ratio/high_mean": 0.0010025296189269284, "clip_ratio/low_mean": 0.0006922776649541144, "clip_ratio/low_min": 0.0001670938749157358, "clip_ratio/region_mean": 0.00169480728606383, "completions/clipped_ratio": 0.0, "completions/max_length": 558.3, "completions/max_terminated_length": 558.3, "completions/mean_length": 197.00859985351562, "completions/mean_terminated_length": 197.00859985351562, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.0263808009835581, "epoch": 1.9464340373431281, "frac_reward_zero_std": 0.16666667275130748, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 325816646.0, "reward": 8.024689769744873, "reward_std": 0.3683147892355919, "rewards/combined_reward/mean": 8.024689483642579, "rewards/combined_reward/std": 2.1221420645713804, "step": 3180 }, { "clip_ratio/high_max": 0.0017751065611567658, "clip_ratio/high_mean": 0.001079682243289426, "clip_ratio/low_mean": 0.0009366761705678073, "clip_ratio/low_min": 0.0002924173010493784, "clip_ratio/region_mean": 0.002016358383601376, "completions/clipped_ratio": 0.0, "completions/max_length": 599.4, "completions/max_terminated_length": 599.4, "completions/mean_length": 208.24271545410156, "completions/mean_terminated_length": 208.24271545410156, "completions/min_length": 48.5, "completions/min_terminated_length": 48.5, "entropy": 0.027973442260796826, "epoch": 1.9647995102540556, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 328916458.0, "reward": 8.002681732177734, "reward_std": 0.3805461645126343, "rewards/combined_reward/mean": 8.002681541442872, "rewards/combined_reward/std": 2.2499295592308046, "step": 3210 }, { "clip_ratio/high_max": 0.0017147687069761256, "clip_ratio/high_mean": 0.000952274245598043, "clip_ratio/low_mean": 0.0007555652504379395, "clip_ratio/low_min": 0.0002137487164873164, "clip_ratio/region_mean": 0.0017078394691149394, "completions/clipped_ratio": 0.0, "completions/max_length": 595.7, "completions/max_terminated_length": 595.7, "completions/mean_length": 200.5851608276367, "completions/mean_terminated_length": 200.5851608276367, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02890395385523637, "epoch": 1.9831649831649831, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.00677490234375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 331986865.0, "reward": 8.213367557525634, "reward_std": 0.3275760367512703, "rewards/combined_reward/mean": 8.213367319107055, "rewards/combined_reward/std": 2.090649056434631, "step": 3240 }, { "clip_ratio/high_max": 0.0019927529912045013, "clip_ratio/high_mean": 0.001081771818924787, "clip_ratio/low_mean": 0.0006654889125531524, "clip_ratio/low_min": 0.00014588015070986614, "clip_ratio/region_mean": 0.0017472607141826302, "completions/clipped_ratio": 0.0, "completions/max_length": 682.5454545454545, "completions/max_terminated_length": 682.5454545454545, "completions/mean_length": 207.63755243474787, "completions/mean_terminated_length": 207.63755243474787, "completions/min_length": 39.09090909090909, "completions/min_terminated_length": 39.09090909090909, "entropy": 0.02804963066257903, "epoch": 2.001836547291093, "frac_reward_zero_std": 0.17803030765869401, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 335394102.0, "reward": 7.902584292671897, "reward_std": 0.4051822302016345, "rewards/combined_reward/mean": 7.902584032578901, "rewards/combined_reward/std": 2.341368458487771, "step": 3270 }, { "clip_ratio/high_max": 0.0025166300576529464, "clip_ratio/high_mean": 0.0014276189568287615, "clip_ratio/low_mean": 0.0010225240004122802, "clip_ratio/low_min": 0.00027207912353333086, "clip_ratio/region_mean": 0.0024501429172232745, "completions/clipped_ratio": 0.0, "completions/max_length": 553.3, "completions/max_terminated_length": 553.3, "completions/mean_length": 198.71849365234374, "completions/mean_terminated_length": 198.71849365234374, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02896855551128586, "epoch": 2.0202020202020203, "frac_reward_zero_std": 0.12500000260770322, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 338457341.0, "reward": 8.286377763748169, "reward_std": 0.3762471854686737, "rewards/combined_reward/mean": 8.286377477645875, "rewards/combined_reward/std": 2.0949577689170837, "step": 3300 }, { "clip_ratio/high_max": 0.002693174326365503, "clip_ratio/high_mean": 0.001508667732802375, "clip_ratio/low_mean": 0.0009580300065863412, "clip_ratio/low_min": 0.000265178740422319, "clip_ratio/region_mean": 0.002466697737690993, "completions/clipped_ratio": 0.0, "completions/max_length": 589.8, "completions/max_terminated_length": 589.8, "completions/mean_length": 201.3424530029297, "completions/mean_terminated_length": 201.3424530029297, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02871395682450384, "epoch": 2.038567493112948, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.01446533203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 341530656.0, "reward": 7.894710206985474, "reward_std": 0.4032824844121933, "rewards/combined_reward/mean": 7.894709825515747, "rewards/combined_reward/std": 2.4488557100296022, "step": 3330 }, { "clip_ratio/high_max": 0.002489988690164561, "clip_ratio/high_mean": 0.0013956904970958324, "clip_ratio/low_mean": 0.001013772955169164, "clip_ratio/low_min": 0.0002573671435432819, "clip_ratio/region_mean": 0.00240946344420081, "completions/clipped_ratio": 0.0, "completions/max_length": 546.5, "completions/max_terminated_length": 546.5, "completions/mean_length": 196.0929748535156, "completions/mean_terminated_length": 196.0929748535156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03018696093931794, "epoch": 2.0569329660238753, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.037109375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 344583813.0, "reward": 8.136496543884277, "reward_std": 0.36816499307751654, "rewards/combined_reward/mean": 8.13649640083313, "rewards/combined_reward/std": 2.1227461457252503, "step": 3360 }, { "clip_ratio/high_max": 0.002490277125131494, "clip_ratio/high_mean": 0.0013114687045648074, "clip_ratio/low_mean": 0.0009759797678270843, "clip_ratio/low_min": 0.00027540609808056614, "clip_ratio/region_mean": 0.0022874484658435296, "completions/clipped_ratio": 0.0, "completions/max_length": 605.4, "completions/max_terminated_length": 605.4, "completions/mean_length": 206.71198425292968, "completions/mean_terminated_length": 206.71198425292968, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.02826305984053761, "epoch": 2.0752984389348024, "frac_reward_zero_std": 0.17916666828095912, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 347677747.0, "reward": 8.125573539733887, "reward_std": 0.3279752418398857, "rewards/combined_reward/mean": 8.125573301315308, "rewards/combined_reward/std": 2.2249075412750243, "step": 3390 }, { "clip_ratio/high_max": 0.002583617068012245, "clip_ratio/high_mean": 0.001398750227963319, "clip_ratio/low_mean": 0.0006918355897748067, "clip_ratio/low_min": 0.0002123758878345446, "clip_ratio/region_mean": 0.002090585787179104, "completions/clipped_ratio": 0.0, "completions/max_length": 598.2, "completions/max_terminated_length": 598.2, "completions/mean_length": 204.10703735351564, "completions/mean_terminated_length": 204.10703735351564, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02656865194439888, "epoch": 2.09366391184573, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 350761678.0, "reward": 8.028842973709107, "reward_std": 0.3488570600748062, "rewards/combined_reward/mean": 8.028842639923095, "rewards/combined_reward/std": 2.4322837471961973, "step": 3420 }, { "clip_ratio/high_max": 0.0024088772086543033, "clip_ratio/high_mean": 0.001331676501528515, "clip_ratio/low_mean": 0.0009456136095347271, "clip_ratio/low_min": 0.0003028110947828585, "clip_ratio/region_mean": 0.002277290090569295, "completions/clipped_ratio": 0.0, "completions/max_length": 708.2, "completions/max_terminated_length": 708.2, "completions/mean_length": 215.45287017822267, "completions/mean_terminated_length": 215.45287017822267, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.032736598917593555, "epoch": 2.1120293847566574, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 353889177.0, "reward": 8.170891189575196, "reward_std": 0.3804438576102257, "rewards/combined_reward/mean": 8.170890998840331, "rewards/combined_reward/std": 2.0776883721351624, "step": 3450 }, { "clip_ratio/high_max": 0.001961899158292605, "clip_ratio/high_mean": 0.001160383668927049, "clip_ratio/low_mean": 0.0010302484407778442, "clip_ratio/low_min": 0.00025077136961044745, "clip_ratio/region_mean": 0.0021906321092198294, "completions/clipped_ratio": 0.0, "completions/max_length": 630.3, "completions/max_terminated_length": 630.3, "completions/mean_length": 206.93724517822267, "completions/mean_terminated_length": 206.93724517822267, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.03036508004491528, "epoch": 2.130394857667585, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 356983976.0, "reward": 7.938913631439209, "reward_std": 0.34044425636529924, "rewards/combined_reward/mean": 7.93891339302063, "rewards/combined_reward/std": 2.4087077379226685, "step": 3480 }, { "clip_ratio/high_max": 0.002541873755399138, "clip_ratio/high_mean": 0.0014289411607023793, "clip_ratio/low_mean": 0.001071805913791953, "clip_ratio/low_min": 0.00030571311435778625, "clip_ratio/region_mean": 0.0025007470661269813, "completions/clipped_ratio": 0.0, "completions/max_length": 562.2, "completions/max_terminated_length": 562.2, "completions/mean_length": 199.22161712646485, "completions/mean_terminated_length": 199.22161712646485, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.031049250074041388, "epoch": 2.1487603305785123, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 360049147.0, "reward": 8.136260843276977, "reward_std": 0.34718275889754296, "rewards/combined_reward/mean": 8.13626070022583, "rewards/combined_reward/std": 2.085802412033081, "step": 3510 }, { "clip_ratio/high_max": 0.002505563533971629, "clip_ratio/high_mean": 0.0014658873821948267, "clip_ratio/low_mean": 0.0007844961795247704, "clip_ratio/low_min": 0.00021051630246802234, "clip_ratio/region_mean": 0.002250383573603661, "completions/clipped_ratio": 0.0, "completions/max_length": 549.9, "completions/max_terminated_length": 549.9, "completions/mean_length": 209.89349517822265, "completions/mean_terminated_length": 209.89349517822265, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.029254214155177274, "epoch": 2.16712580348944, "frac_reward_zero_std": 0.15000000558793544, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 363155298.0, "reward": 8.260387849807739, "reward_std": 0.3202134594321251, "rewards/combined_reward/mean": 8.260387706756593, "rewards/combined_reward/std": 2.048438048362732, "step": 3540 }, { "clip_ratio/high_max": 0.002423270620541492, "clip_ratio/high_mean": 0.001437063614988195, "clip_ratio/low_mean": 0.001035562073229812, "clip_ratio/low_min": 0.0003221146844831916, "clip_ratio/region_mean": 0.002472625684943826, "completions/clipped_ratio": 0.0, "completions/max_length": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 202.66693420410155, "completions/mean_terminated_length": 202.66693420410155, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.029769730187642077, "epoch": 2.1854912764003673, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 366233699.0, "reward": 7.975155448913574, "reward_std": 0.39104330390691755, "rewards/combined_reward/mean": 7.975155162811279, "rewards/combined_reward/std": 2.460030663013458, "step": 3570 }, { "clip_ratio/high_max": 0.002436797381475723, "clip_ratio/high_mean": 0.0013756264942154908, "clip_ratio/low_mean": 0.0010250241599957615, "clip_ratio/low_min": 0.0002741483639207824, "clip_ratio/region_mean": 0.0024006506690056996, "completions/clipped_ratio": 0.0, "completions/max_length": 605.9, "completions/max_terminated_length": 605.9, "completions/mean_length": 195.0226593017578, "completions/mean_terminated_length": 195.0226593017578, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03080531454955538, "epoch": 2.203856749311295, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 369282746.0, "reward": 8.21471357345581, "reward_std": 0.334880081564188, "rewards/combined_reward/mean": 8.214713144302369, "rewards/combined_reward/std": 2.2087485671043394, "step": 3600 }, { "clip_ratio/high_max": 0.0022391820001454715, "clip_ratio/high_mean": 0.0012615760458478083, "clip_ratio/low_mean": 0.0009559492141609856, "clip_ratio/low_min": 0.0003085582354591073, "clip_ratio/region_mean": 0.002217525263646773, "completions/clipped_ratio": 0.0, "completions/max_length": 643.7, "completions/max_terminated_length": 643.7, "completions/mean_length": 198.13490142822266, "completions/mean_terminated_length": 198.13490142822266, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.027988295381267864, "epoch": 2.2222222222222223, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 372343744.0, "reward": 7.94391508102417, "reward_std": 0.40710687041282656, "rewards/combined_reward/mean": 7.943914604187012, "rewards/combined_reward/std": 2.3980599999427796, "step": 3630 }, { "clip_ratio/high_max": 0.0020980893294715013, "clip_ratio/high_mean": 0.0012287236653113116, "clip_ratio/low_mean": 0.0010371457816897116, "clip_ratio/low_min": 0.0003350729743639628, "clip_ratio/region_mean": 0.0022658694579149596, "completions/clipped_ratio": 0.0, "completions/max_length": 588.6, "completions/max_terminated_length": 588.6, "completions/mean_length": 198.77240142822265, "completions/mean_terminated_length": 198.77240142822265, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02872112885427972, "epoch": 2.24058769513315, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 375407190.0, "reward": 7.968915176391602, "reward_std": 0.3165806412696838, "rewards/combined_reward/mean": 7.968915128707886, "rewards/combined_reward/std": 2.179728937149048, "step": 3660 }, { "clip_ratio/high_max": 0.002766316402509498, "clip_ratio/high_mean": 0.0015991927136686475, "clip_ratio/low_mean": 0.0009260675509115875, "clip_ratio/low_min": 0.0002788423293774637, "clip_ratio/region_mean": 0.0025252602911981135, "completions/clipped_ratio": 0.0, "completions/max_length": 609.7, "completions/max_terminated_length": 609.7, "completions/mean_length": 203.0796920776367, "completions/mean_terminated_length": 203.0796920776367, "completions/min_length": 45.3, "completions/min_terminated_length": 45.3, "entropy": 0.03510390729643405, "epoch": 2.2589531680440773, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0033, "num_tokens": 378487176.0, "reward": 8.019264221191406, "reward_std": 0.4042950063943863, "rewards/combined_reward/mean": 8.019263553619385, "rewards/combined_reward/std": 2.308472156524658, "step": 3690 }, { "clip_ratio/high_max": 0.0027296289641526527, "clip_ratio/high_mean": 0.001619274889647689, "clip_ratio/low_mean": 0.000953010198524377, "clip_ratio/low_min": 0.00024148105051911747, "clip_ratio/region_mean": 0.0025722851297662903, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 197.41615295410156, "completions/mean_terminated_length": 197.41615295410156, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03141545240456859, "epoch": 2.277318640955005, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 381545414.0, "reward": 8.155217027664184, "reward_std": 0.36707631051540374, "rewards/combined_reward/mean": 8.155217027664184, "rewards/combined_reward/std": 2.0539712429046633, "step": 3720 }, { "clip_ratio/high_max": 0.002358835497094939, "clip_ratio/high_mean": 0.001374785858206451, "clip_ratio/low_mean": 0.0008960614640576144, "clip_ratio/low_min": 0.000286552008765284, "clip_ratio/region_mean": 0.002270847325902044, "completions/clipped_ratio": 0.0, "completions/max_length": 622.7, "completions/max_terminated_length": 622.7, "completions/mean_length": 196.4830795288086, "completions/mean_terminated_length": 196.4830795288086, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.0280869516544044, "epoch": 2.2956841138659323, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.0123291015625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 384600069.0, "reward": 7.998271894454956, "reward_std": 0.3628596544265747, "rewards/combined_reward/mean": 7.998271656036377, "rewards/combined_reward/std": 2.3294278383255005, "step": 3750 }, { "clip_ratio/high_max": 0.003669811507764583, "clip_ratio/high_mean": 0.0020315852074418218, "clip_ratio/low_mean": 0.001042106224304007, "clip_ratio/low_min": 0.0003199762463433823, "clip_ratio/region_mean": 0.003073691446722175, "completions/clipped_ratio": 0.0, "completions/max_length": 670.5, "completions/max_terminated_length": 670.5, "completions/mean_length": 195.92526397705078, "completions/mean_terminated_length": 195.92526397705078, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.041749778545151155, "epoch": 2.3140495867768593, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 387652582.0, "reward": 7.979736280441284, "reward_std": 0.3632249146699905, "rewards/combined_reward/mean": 7.97973599433899, "rewards/combined_reward/std": 2.3784743309020997, "step": 3780 }, { "clip_ratio/high_max": 0.002552892966681005, "clip_ratio/high_mean": 0.0014680382104415913, "clip_ratio/low_mean": 0.00096280867534612, "clip_ratio/low_min": 0.00026277493549666057, "clip_ratio/region_mean": 0.002430846874873775, "completions/clipped_ratio": 0.0, "completions/max_length": 585.9, "completions/max_terminated_length": 585.9, "completions/mean_length": 201.2671920776367, "completions/mean_terminated_length": 201.2671920776367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03123981667061647, "epoch": 2.332415059687787, "frac_reward_zero_std": 0.14166666902601718, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 390725608.0, "reward": 8.030435943603516, "reward_std": 0.37510950267314913, "rewards/combined_reward/mean": 8.030435752868652, "rewards/combined_reward/std": 2.2813759326934813, "step": 3810 }, { "clip_ratio/high_max": 0.0024978938245718988, "clip_ratio/high_mean": 0.001439475365987164, "clip_ratio/low_mean": 0.001159840976833948, "clip_ratio/low_min": 0.0003530290191217015, "clip_ratio/region_mean": 0.0025993163416084524, "completions/clipped_ratio": 0.0, "completions/max_length": 769.4, "completions/max_terminated_length": 769.4, "completions/mean_length": 198.42630920410156, "completions/mean_terminated_length": 198.42630920410156, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.029632447116697828, "epoch": 2.3507805325987143, "frac_reward_zero_std": 0.18333333656191825, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 393787725.0, "reward": 8.206747245788574, "reward_std": 0.25879030376672746, "rewards/combined_reward/mean": 8.206747055053711, "rewards/combined_reward/std": 2.110676896572113, "step": 3840 }, { "clip_ratio/high_max": 0.002360222799567661, "clip_ratio/high_mean": 0.0013045034416184838, "clip_ratio/low_mean": 0.0008913551183771536, "clip_ratio/low_min": 0.00027807726049407694, "clip_ratio/region_mean": 0.0021958585629666535, "completions/clipped_ratio": 0.0, "completions/max_length": 593.2, "completions/max_terminated_length": 593.2, "completions/mean_length": 200.99557800292968, "completions/mean_terminated_length": 200.99557800292968, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.028252803285916648, "epoch": 2.369146005509642, "frac_reward_zero_std": 0.18750000298023223, "grad_norm": 0.017822265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 396859708.0, "reward": 8.181289386749267, "reward_std": 0.3424511641263962, "rewards/combined_reward/mean": 8.181289196014404, "rewards/combined_reward/std": 2.204875981807709, "step": 3870 }, { "clip_ratio/high_max": 0.0021545095897939365, "clip_ratio/high_mean": 0.0013046611855922188, "clip_ratio/low_mean": 0.0009430244360070598, "clip_ratio/low_min": 0.0002815116189594846, "clip_ratio/region_mean": 0.0022476856160210446, "completions/clipped_ratio": 0.0, "completions/max_length": 510.4, "completions/max_terminated_length": 510.4, "completions/mean_length": 191.16250610351562, "completions/mean_terminated_length": 191.16250610351562, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.02872219371298949, "epoch": 2.3875114784205693, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 399893932.0, "reward": 8.236712837219239, "reward_std": 0.2568485766649246, "rewards/combined_reward/mean": 8.23671259880066, "rewards/combined_reward/std": 2.137237215042114, "step": 3900 }, { "clip_ratio/high_max": 0.0023441650133463553, "clip_ratio/high_mean": 0.0013306869164807721, "clip_ratio/low_mean": 0.000881459148755918, "clip_ratio/low_min": 0.00019360365671066876, "clip_ratio/region_mean": 0.0022121460404984344, "completions/clipped_ratio": 0.0, "completions/max_length": 570.1, "completions/max_terminated_length": 570.1, "completions/mean_length": 188.80703735351562, "completions/mean_terminated_length": 188.80703735351562, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.026875358525042733, "epoch": 2.405876951331497, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.00701904296875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 402919111.0, "reward": 8.310951328277588, "reward_std": 0.31362170577049253, "rewards/combined_reward/mean": 8.310951471328735, "rewards/combined_reward/std": 1.8243084311485291, "step": 3930 }, { "clip_ratio/high_max": 0.0026771602574929905, "clip_ratio/high_mean": 0.0014340949136870526, "clip_ratio/low_mean": 0.0009496337492843547, "clip_ratio/low_min": 0.00021908864388630415, "clip_ratio/region_mean": 0.00238372866588179, "completions/clipped_ratio": 0.0, "completions/max_length": 656.7, "completions/max_terminated_length": 656.7, "completions/mean_length": 199.7072952270508, "completions/mean_terminated_length": 199.7072952270508, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.030601398763246836, "epoch": 2.4242424242424243, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 405986147.0, "reward": 7.850231742858886, "reward_std": 0.3338914915919304, "rewards/combined_reward/mean": 7.850231599807739, "rewards/combined_reward/std": 2.567798936367035, "step": 3960 }, { "clip_ratio/high_max": 0.002461528482672293, "clip_ratio/high_mean": 0.0013929110250804418, "clip_ratio/low_mean": 0.0009681850357689351, "clip_ratio/low_min": 0.00022698197159722137, "clip_ratio/region_mean": 0.002361096078918005, "completions/clipped_ratio": 0.0, "completions/max_length": 565.5, "completions/max_terminated_length": 565.5, "completions/mean_length": 197.10131072998047, "completions/mean_terminated_length": 197.10131072998047, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.032221662299707535, "epoch": 2.442607897153352, "frac_reward_zero_std": 0.1708333410322666, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 409043176.0, "reward": 7.9711884498596195, "reward_std": 0.3386555194854736, "rewards/combined_reward/mean": 7.9711884498596195, "rewards/combined_reward/std": 2.439654541015625, "step": 3990 }, { "clip_ratio/high_max": 0.0023931161529617386, "clip_ratio/high_mean": 0.0013305864011878535, "clip_ratio/low_mean": 0.0010173324460386842, "clip_ratio/low_min": 0.00026566820136698276, "clip_ratio/region_mean": 0.00234791884116324, "completions/clipped_ratio": 0.0, "completions/max_length": 604.2, "completions/max_terminated_length": 604.2, "completions/mean_length": 195.5690170288086, "completions/mean_terminated_length": 195.5690170288086, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.030339051208769282, "epoch": 2.4609733700642793, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 412094321.0, "reward": 7.877132272720337, "reward_std": 0.34861014783382416, "rewards/combined_reward/mean": 7.877132177352905, "rewards/combined_reward/std": 2.444123721122742, "step": 4020 }, { "clip_ratio/high_max": 0.002563158390209234, "clip_ratio/high_mean": 0.0014303796184928312, "clip_ratio/low_mean": 0.0008951120853453176, "clip_ratio/low_min": 0.0002652853584246865, "clip_ratio/region_mean": 0.002325491709295117, "completions/clipped_ratio": 0.0, "completions/max_length": 573.3, "completions/max_terminated_length": 573.3, "completions/mean_length": 201.04870147705077, "completions/mean_terminated_length": 201.04870147705077, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.029340796684846283, "epoch": 2.479338842975207, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 415166508.0, "reward": 8.170740938186645, "reward_std": 0.34740202575922013, "rewards/combined_reward/mean": 8.170740461349487, "rewards/combined_reward/std": 2.3597180247306824, "step": 4050 }, { "clip_ratio/high_max": 0.002622246178604352, "clip_ratio/high_mean": 0.0013020679956146826, "clip_ratio/low_mean": 0.000896467150838968, "clip_ratio/low_min": 0.00024929457576945425, "clip_ratio/region_mean": 0.00219853513117414, "completions/clipped_ratio": 0.0, "completions/max_length": 526.7, "completions/max_terminated_length": 526.7, "completions/mean_length": 185.04193115234375, "completions/mean_terminated_length": 185.04193115234375, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.029495826037600637, "epoch": 2.497704315886134, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 418177229.0, "reward": 7.977499675750733, "reward_std": 0.32838964760303496, "rewards/combined_reward/mean": 7.9774994373321535, "rewards/combined_reward/std": 2.36697803735733, "step": 4080 }, { "clip_ratio/high_max": 0.0027862690015657185, "clip_ratio/high_mean": 0.0015902741387738692, "clip_ratio/low_mean": 0.001092756765926121, "clip_ratio/low_min": 0.00039665710355620833, "clip_ratio/region_mean": 0.0026830308876621225, "completions/clipped_ratio": 0.0, "completions/max_length": 606.7, "completions/max_terminated_length": 606.7, "completions/mean_length": 201.5768310546875, "completions/mean_terminated_length": 201.5768310546875, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "entropy": 0.033316551707684994, "epoch": 2.5160697887970613, "frac_reward_zero_std": 0.15000000298023225, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 421251444.0, "reward": 7.89650936126709, "reward_std": 0.2551050387322903, "rewards/combined_reward/mean": 7.896509218215942, "rewards/combined_reward/std": 2.5333843350410463, "step": 4110 }, { "clip_ratio/high_max": 0.0023471250092067446, "clip_ratio/high_mean": 0.0013708348085250085, "clip_ratio/low_mean": 0.001086706384497423, "clip_ratio/low_min": 0.0003025920838505651, "clip_ratio/region_mean": 0.0024575411870803995, "completions/clipped_ratio": 0.0, "completions/max_length": 555.6, "completions/max_terminated_length": 555.6, "completions/mean_length": 197.50599365234376, "completions/mean_terminated_length": 197.50599365234376, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.032584662875160576, "epoch": 2.534435261707989, "frac_reward_zero_std": 0.1833333358168602, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 424310027.0, "reward": 8.271001243591309, "reward_std": 0.3126299291849136, "rewards/combined_reward/mean": 8.27100067138672, "rewards/combined_reward/std": 2.010485887527466, "step": 4140 }, { "clip_ratio/high_max": 0.0023222427824900176, "clip_ratio/high_mean": 0.0014030250184684215, "clip_ratio/low_mean": 0.0008816693186721144, "clip_ratio/low_min": 0.00025624690776263986, "clip_ratio/region_mean": 0.002284694342233706, "completions/clipped_ratio": 0.0, "completions/max_length": 653.8, "completions/max_terminated_length": 653.8, "completions/mean_length": 205.38620300292968, "completions/mean_terminated_length": 205.38620300292968, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0280533398501575, "epoch": 2.5528007346189163, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 427398870.0, "reward": 8.009939098358155, "reward_std": 0.40968642979860304, "rewards/combined_reward/mean": 8.009938621520996, "rewards/combined_reward/std": 2.2861735463142394, "step": 4170 }, { "clip_ratio/high_max": 0.0031736995433069145, "clip_ratio/high_mean": 0.0016640623556061958, "clip_ratio/low_mean": 0.0011315125651890413, "clip_ratio/low_min": 0.0003269615714089014, "clip_ratio/region_mean": 0.002795574933406897, "completions/clipped_ratio": 0.0, "completions/max_length": 633.1, "completions/max_terminated_length": 633.1, "completions/mean_length": 202.1994857788086, "completions/mean_terminated_length": 202.1994857788086, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.031878946473201115, "epoch": 2.571166207529844, "frac_reward_zero_std": 0.1916666742414236, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 430475476.0, "reward": 8.047818422317505, "reward_std": 0.363895520567894, "rewards/combined_reward/mean": 8.047818374633788, "rewards/combined_reward/std": 2.1167474746704102, "step": 4200 }, { "clip_ratio/high_max": 0.002410870231688023, "clip_ratio/high_mean": 0.0013947298073617275, "clip_ratio/low_mean": 0.0011307644092691286, "clip_ratio/low_min": 0.00029744811763521284, "clip_ratio/region_mean": 0.002525494207899707, "completions/clipped_ratio": 0.0, "completions/max_length": 607.3, "completions/max_terminated_length": 607.3, "completions/mean_length": 200.69818267822265, "completions/mean_terminated_length": 200.69818267822265, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.02973576933145523, "epoch": 2.5895316804407713, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 433546317.0, "reward": 8.138554954528809, "reward_std": 0.33838226199150084, "rewards/combined_reward/mean": 8.13855471611023, "rewards/combined_reward/std": 2.226789343357086, "step": 4230 }, { "clip_ratio/high_max": 0.0023245943229994737, "clip_ratio/high_mean": 0.0013566038454882801, "clip_ratio/low_mean": 0.0010090009003761225, "clip_ratio/low_min": 0.0003356492338449849, "clip_ratio/region_mean": 0.0023656047676922755, "completions/clipped_ratio": 0.0, "completions/max_length": 603.6, "completions/max_terminated_length": 603.6, "completions/mean_length": 204.6268280029297, "completions/mean_terminated_length": 204.6268280029297, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026823841120737293, "epoch": 2.607897153351699, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.023681640625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 436632244.0, "reward": 8.288584089279174, "reward_std": 0.3608225703239441, "rewards/combined_reward/mean": 8.288583517074585, "rewards/combined_reward/std": 1.987896740436554, "step": 4260 }, { "clip_ratio/high_max": 0.0027352509525371714, "clip_ratio/high_mean": 0.001620294431268121, "clip_ratio/low_mean": 0.0010454106294976858, "clip_ratio/low_min": 0.00027123798742347086, "clip_ratio/region_mean": 0.002665705048032881, "completions/clipped_ratio": 0.0, "completions/max_length": 595.6, "completions/max_terminated_length": 595.6, "completions/mean_length": 199.30312957763672, "completions/mean_terminated_length": 199.30312957763672, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.030857501458376645, "epoch": 2.6262626262626263, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 439697728.0, "reward": 7.819858837127685, "reward_std": 0.3313475251197815, "rewards/combined_reward/mean": 7.819858407974243, "rewards/combined_reward/std": 2.411305582523346, "step": 4290 }, { "clip_ratio/high_max": 0.002390074392315, "clip_ratio/high_mean": 0.001392941493395483, "clip_ratio/low_mean": 0.0010133768350594134, "clip_ratio/low_min": 0.0003210800830856897, "clip_ratio/region_mean": 0.002406318332335407, "completions/clipped_ratio": 0.0, "completions/max_length": 592.6, "completions/max_terminated_length": 592.6, "completions/mean_length": 196.5031311035156, "completions/mean_terminated_length": 196.5031311035156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03044891352765262, "epoch": 2.644628099173554, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 442752460.0, "reward": 7.835739088058472, "reward_std": 0.36730786561965945, "rewards/combined_reward/mean": 7.835738897323608, "rewards/combined_reward/std": 2.380213975906372, "step": 4320 }, { "clip_ratio/high_max": 0.0022531277892994693, "clip_ratio/high_mean": 0.001334424118855774, "clip_ratio/low_mean": 0.000993518760757676, "clip_ratio/low_min": 0.0003641230643552262, "clip_ratio/region_mean": 0.00232794288021978, "completions/clipped_ratio": 0.0, "completions/max_length": 506.1, "completions/max_terminated_length": 506.1, "completions/mean_length": 189.27708740234374, "completions/mean_terminated_length": 189.27708740234374, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.028446434965978065, "epoch": 2.6629935720844813, "frac_reward_zero_std": 0.16666667424142362, "grad_norm": 0.01446533203125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 445779444.0, "reward": 8.127084112167358, "reward_std": 0.29514688402414324, "rewards/combined_reward/mean": 8.127083778381348, "rewards/combined_reward/std": 2.189292287826538, "step": 4350 }, { "clip_ratio/high_max": 0.0023400386460707525, "clip_ratio/high_mean": 0.0014163119104826669, "clip_ratio/low_mean": 0.000986383297640714, "clip_ratio/low_min": 0.0002462664987736692, "clip_ratio/region_mean": 0.002402695206789455, "completions/clipped_ratio": 0.0, "completions/max_length": 702.1, "completions/max_terminated_length": 702.1, "completions/mean_length": 216.8830764770508, "completions/mean_terminated_length": 216.8830764770508, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.02929143194730083, "epoch": 2.681359044995409, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 448912435.0, "reward": 7.958446741104126, "reward_std": 0.3850564658641815, "rewards/combined_reward/mean": 7.9584465503692625, "rewards/combined_reward/std": 2.4539627194404603, "step": 4380 }, { "clip_ratio/high_max": 0.0020295763853937386, "clip_ratio/high_mean": 0.001207527227961691, "clip_ratio/low_mean": 0.0008505003657774069, "clip_ratio/low_min": 0.0002830700839695055, "clip_ratio/region_mean": 0.0020580275915563105, "completions/clipped_ratio": 0.0, "completions/max_length": 831.75, "completions/max_terminated_length": 831.75, "completions/mean_length": 196.8431053161621, "completions/mean_terminated_length": 196.8431053161621, "completions/min_length": 37.25, "completions/min_terminated_length": 37.25, "entropy": 0.029589621955528854, "epoch": 2.699112335475972, "frac_reward_zero_std": 0.1770833358168602, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 451970625.0, "reward": 7.733312487602234, "reward_std": 0.39512546360492706, "rewards/combined_reward/mean": 7.733312249183655, "rewards/combined_reward/std": 2.6765098571777344, "step": 4410 }, { "clip_ratio/high_max": 0.0019385022188847265, "clip_ratio/high_mean": 0.001127429893798156, "clip_ratio/low_mean": 0.000771301000471188, "clip_ratio/low_min": 0.0001721750020806212, "clip_ratio/region_mean": 0.0018987308896612376, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 206.3354248046875, "completions/mean_terminated_length": 206.3354248046875, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.027220508370858928, "epoch": 2.7174778083868993, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 455063113.0, "reward": 7.876735401153565, "reward_std": 0.3926775738596916, "rewards/combined_reward/mean": 7.876734972000122, "rewards/combined_reward/std": 2.4638564586639404, "step": 4440 }, { "clip_ratio/high_max": 0.0022597033452863495, "clip_ratio/high_mean": 0.0012963331074085241, "clip_ratio/low_mean": 0.0008396926855008739, "clip_ratio/low_min": 0.0002713687703362666, "clip_ratio/region_mean": 0.0021360257892714192, "completions/clipped_ratio": 0.0, "completions/max_length": 544.4, "completions/max_terminated_length": 544.4, "completions/mean_length": 199.7291732788086, "completions/mean_terminated_length": 199.7291732788086, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.030568143376149236, "epoch": 2.735843281297827, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 458130233.0, "reward": 7.996156024932861, "reward_std": 0.3222914829850197, "rewards/combined_reward/mean": 7.996155691146851, "rewards/combined_reward/std": 2.410098648071289, "step": 4470 }, { "clip_ratio/high_max": 0.002224176423624158, "clip_ratio/high_mean": 0.0012373582908670262, "clip_ratio/low_mean": 0.0007879497544005669, "clip_ratio/low_min": 0.00023337541667084832, "clip_ratio/region_mean": 0.0020253080467227845, "completions/clipped_ratio": 0.0, "completions/max_length": 623.3, "completions/max_terminated_length": 623.3, "completions/mean_length": 206.8783920288086, "completions/mean_terminated_length": 206.8783920288086, "completions/min_length": 51.8, "completions/min_terminated_length": 51.8, "entropy": 0.03208857782495519, "epoch": 2.7542087542087543, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 461224806.0, "reward": 8.0490797996521, "reward_std": 0.33945527821779253, "rewards/combined_reward/mean": 8.04907956123352, "rewards/combined_reward/std": 2.317999517917633, "step": 4500 }, { "clip_ratio/high_max": 0.0016035159598686733, "clip_ratio/high_mean": 0.0009993048353256504, "clip_ratio/low_mean": 0.0006388450293646504, "clip_ratio/low_min": 0.00014339169259377135, "clip_ratio/region_mean": 0.0016381498586270026, "completions/clipped_ratio": 0.0, "completions/max_length": 579.9, "completions/max_terminated_length": 579.9, "completions/mean_length": 196.85104522705078, "completions/mean_terminated_length": 196.85104522705078, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.02672720330301672, "epoch": 2.772574227119682, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.007537841796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 464280874.0, "reward": 8.366993045806884, "reward_std": 0.29280704706907273, "rewards/combined_reward/mean": 8.36699275970459, "rewards/combined_reward/std": 1.984869933128357, "step": 4530 }, { "clip_ratio/high_max": 0.0016736330318963156, "clip_ratio/high_mean": 0.0009269931989062268, "clip_ratio/low_mean": 0.0007949272302842776, "clip_ratio/low_min": 0.00023084051378342944, "clip_ratio/region_mean": 0.0017219204227634085, "completions/clipped_ratio": 0.0, "completions/max_length": 571.1, "completions/max_terminated_length": 571.1, "completions/mean_length": 210.1002670288086, "completions/mean_terminated_length": 210.1002670288086, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.028789468730489414, "epoch": 2.7909397000306093, "frac_reward_zero_std": 0.1333333358168602, "grad_norm": 0.005767822265625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 467387819.0, "reward": 8.166336250305175, "reward_std": 0.34855775237083436, "rewards/combined_reward/mean": 8.166335964202881, "rewards/combined_reward/std": 2.1776638865470885, "step": 4560 }, { "clip_ratio/high_max": 0.0021700367268446524, "clip_ratio/high_mean": 0.0012579514794197166, "clip_ratio/low_mean": 0.0007881012563908977, "clip_ratio/low_min": 0.00021363533475475074, "clip_ratio/region_mean": 0.0020460527167112256, "completions/clipped_ratio": 0.0, "completions/max_length": 680.1, "completions/max_terminated_length": 680.1, "completions/mean_length": 203.32474517822266, "completions/mean_terminated_length": 203.32474517822266, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.03379311376872162, "epoch": 2.809305172941537, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0250244140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 470468746.0, "reward": 7.800893259048462, "reward_std": 0.3415580362081528, "rewards/combined_reward/mean": 7.800892972946167, "rewards/combined_reward/std": 2.4637797594070436, "step": 4590 }, { "clip_ratio/high_max": 0.0021597624256780063, "clip_ratio/high_mean": 0.001224200637686105, "clip_ratio/low_mean": 0.0007274423641016862, "clip_ratio/low_min": 0.00022350421228717703, "clip_ratio/region_mean": 0.0019516430038493127, "completions/clipped_ratio": 0.0, "completions/max_length": 624.2, "completions/max_terminated_length": 624.2, "completions/mean_length": 206.11328582763673, "completions/mean_terminated_length": 206.11328582763673, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.032156545156612995, "epoch": 2.8276706458524643, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 473560381.0, "reward": 7.977965593338013, "reward_std": 0.3960563033819199, "rewards/combined_reward/mean": 7.977965307235718, "rewards/combined_reward/std": 2.203814387321472, "step": 4620 }, { "clip_ratio/high_max": 0.0018743267636940194, "clip_ratio/high_mean": 0.001060224300211606, "clip_ratio/low_mean": 0.0007373056305368664, "clip_ratio/low_min": 0.00018573525060977165, "clip_ratio/region_mean": 0.001797529925534036, "completions/clipped_ratio": 0.0, "completions/max_length": 492.4, "completions/max_terminated_length": 492.4, "completions/mean_length": 187.2487045288086, "completions/mean_terminated_length": 187.2487045288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02652368840451042, "epoch": 2.8460361187633914, "frac_reward_zero_std": 0.18750000223517418, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 476579576.0, "reward": 7.963632583618164, "reward_std": 0.2907959148287773, "rewards/combined_reward/mean": 7.963632249832154, "rewards/combined_reward/std": 2.358007490634918, "step": 4650 }, { "clip_ratio/high_max": 0.0016972513777242662, "clip_ratio/high_mean": 0.0010058029656647705, "clip_ratio/low_mean": 0.0006550427808785268, "clip_ratio/low_min": 0.00013671529571486948, "clip_ratio/region_mean": 0.001660845741086329, "completions/clipped_ratio": 0.0, "completions/max_length": 637.3, "completions/max_terminated_length": 637.3, "completions/mean_length": 204.32500152587892, "completions/mean_terminated_length": 204.32500152587892, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.02569670877419412, "epoch": 2.864401591674319, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 479664344.0, "reward": 8.046384859085084, "reward_std": 0.35275447815656663, "rewards/combined_reward/mean": 8.046384572982788, "rewards/combined_reward/std": 2.438334107398987, "step": 4680 }, { "clip_ratio/high_max": 0.0018642006859105702, "clip_ratio/high_mean": 0.0010819589886523317, "clip_ratio/low_mean": 0.0008032846655017541, "clip_ratio/low_min": 0.00019021230618818664, "clip_ratio/region_mean": 0.0018852436585196604, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 208.04349517822266, "completions/mean_terminated_length": 208.04349517822266, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.033525212993845345, "epoch": 2.8827670645852463, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.00640869140625, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 482763391.0, "reward": 8.105202341079712, "reward_std": 0.4250373363494873, "rewards/combined_reward/mean": 8.105202054977417, "rewards/combined_reward/std": 2.231715202331543, "step": 4710 }, { "clip_ratio/high_max": 0.0021528468050140267, "clip_ratio/high_mean": 0.0011636232139911348, "clip_ratio/low_mean": 0.0008292804974189494, "clip_ratio/low_min": 0.00027844126040387586, "clip_ratio/region_mean": 0.0019929037232941484, "completions/clipped_ratio": 0.0, "completions/max_length": 591.9, "completions/max_terminated_length": 591.9, "completions/mean_length": 206.8908920288086, "completions/mean_terminated_length": 206.8908920288086, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030356007783363264, "epoch": 2.901132537496174, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 485858012.0, "reward": 8.05761022567749, "reward_std": 0.3586785852909088, "rewards/combined_reward/mean": 8.057609939575196, "rewards/combined_reward/std": 2.214602994918823, "step": 4740 }, { "clip_ratio/high_max": 0.0021677939599612727, "clip_ratio/high_mean": 0.0011776779966263954, "clip_ratio/low_mean": 0.0008946159569556282, "clip_ratio/low_min": 0.00029592671586821476, "clip_ratio/region_mean": 0.002072293945457204, "completions/clipped_ratio": 0.0, "completions/max_length": 646.2, "completions/max_terminated_length": 646.2, "completions/mean_length": 212.6117202758789, "completions/mean_terminated_length": 212.6117202758789, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03269302741003533, "epoch": 2.9194980104071013, "frac_reward_zero_std": 0.13333333507180214, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 488974601.0, "reward": 7.702929162979126, "reward_std": 0.4054496169090271, "rewards/combined_reward/mean": 7.702929067611694, "rewards/combined_reward/std": 2.4355583071708677, "step": 4770 }, { "clip_ratio/high_max": 0.0017854200394746536, "clip_ratio/high_mean": 0.0009850230987164348, "clip_ratio/low_mean": 0.0007343938183718516, "clip_ratio/low_min": 0.00021112394533702173, "clip_ratio/region_mean": 0.0017194169250918398, "completions/clipped_ratio": 0.0, "completions/max_length": 546.6, "completions/max_terminated_length": 546.6, "completions/mean_length": 192.07552490234374, "completions/mean_terminated_length": 192.07552490234374, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.02697083642706275, "epoch": 2.937863483318029, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.0223388671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 492012331.0, "reward": 8.269071865081788, "reward_std": 0.2774563401937485, "rewards/combined_reward/mean": 8.269071817398071, "rewards/combined_reward/std": 1.850505495071411, "step": 4800 }, { "clip_ratio/high_max": 0.0018500171790947206, "clip_ratio/high_mean": 0.0010876999028065863, "clip_ratio/low_mean": 0.0006668345303599683, "clip_ratio/low_min": 0.00014533716797207792, "clip_ratio/region_mean": 0.0017545344434135282, "completions/clipped_ratio": 0.0, "completions/max_length": 594.6, "completions/max_terminated_length": 594.6, "completions/mean_length": 219.38281707763673, "completions/mean_terminated_length": 219.38281707763673, "completions/min_length": 48.4, "completions/min_terminated_length": 48.4, "entropy": 0.027947819496815404, "epoch": 2.9562289562289563, "frac_reward_zero_std": 0.13333333805203437, "grad_norm": 0.02685546875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 495154921.0, "reward": 7.877281761169433, "reward_std": 0.3823466882109642, "rewards/combined_reward/mean": 7.877281522750854, "rewards/combined_reward/std": 2.4451729536056517, "step": 4830 }, { "clip_ratio/high_max": 0.0020217924010163795, "clip_ratio/high_mean": 0.0011937687918058752, "clip_ratio/low_mean": 0.0006808226940847817, "clip_ratio/low_min": 0.00015594481155858375, "clip_ratio/region_mean": 0.0018745914882553432, "completions/clipped_ratio": 0.0, "completions/max_length": 580.7, "completions/max_terminated_length": 580.7, "completions/mean_length": 195.5343795776367, "completions/mean_terminated_length": 195.5343795776367, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.027664717038472494, "epoch": 2.974594429139884, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.0277099609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 498205933.0, "reward": 8.225726127624512, "reward_std": 0.3517259329557419, "rewards/combined_reward/mean": 8.22572603225708, "rewards/combined_reward/std": 2.0046374082565306, "step": 4860 }, { "clip_ratio/high_max": 0.0023395312600769103, "clip_ratio/high_mean": 0.0012243167308042756, "clip_ratio/low_mean": 0.0007463305464625591, "clip_ratio/low_min": 0.0001422156749564844, "clip_ratio/region_mean": 0.001970647297275718, "completions/clipped_ratio": 0.0, "completions/max_length": 605.1, "completions/max_terminated_length": 605.1, "completions/mean_length": 206.18438110351562, "completions/mean_terminated_length": 206.18438110351562, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03051868557619552, "epoch": 2.9929599020508113, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 501297841.0, "reward": 8.084622669219971, "reward_std": 0.3237067207694054, "rewards/combined_reward/mean": 8.084622526168824, "rewards/combined_reward/std": 2.212592124938965, "step": 4890 }, { "clip_ratio/high_max": 0.0019266678776103453, "clip_ratio/high_mean": 0.0010887727093192002, "clip_ratio/low_mean": 0.0008383355376341182, "clip_ratio/low_min": 0.00024561718612962756, "clip_ratio/region_mean": 0.001927108247668986, "completions/clipped_ratio": 0.0, "completions/max_length": 636.4, "completions/max_terminated_length": 636.4, "completions/mean_length": 207.76328887939454, "completions/mean_terminated_length": 207.76328887939454, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.025742992224385502, "epoch": 3.0116314661769206, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 504395812.0, "reward": 8.02634539604187, "reward_std": 0.351841402053833, "rewards/combined_reward/mean": 8.026344919204712, "rewards/combined_reward/std": 2.2308682799339294, "step": 4920 }, { "clip_ratio/high_max": 0.0022260128607740626, "clip_ratio/high_mean": 0.0013529595404785748, "clip_ratio/low_mean": 0.001221133317812928, "clip_ratio/low_min": 0.00044088475575942233, "clip_ratio/region_mean": 0.0025740928768451946, "completions/clipped_ratio": 0.0, "completions/max_length": 604.1, "completions/max_terminated_length": 604.1, "completions/mean_length": 207.19114990234374, "completions/mean_terminated_length": 207.19114990234374, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.03074007445635895, "epoch": 3.029996939087848, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 507491586.0, "reward": 8.21140079498291, "reward_std": 0.376400463283062, "rewards/combined_reward/mean": 8.211400556564332, "rewards/combined_reward/std": 2.197994148731232, "step": 4950 }, { "clip_ratio/high_max": 0.002549348569785555, "clip_ratio/high_mean": 0.0013916043055360206, "clip_ratio/low_mean": 0.0008204099661573613, "clip_ratio/low_min": 0.00028185489500174297, "clip_ratio/region_mean": 0.002212014289398212, "completions/clipped_ratio": 0.0, "completions/max_length": 539.7, "completions/max_terminated_length": 539.7, "completions/mean_length": 195.17526550292968, "completions/mean_terminated_length": 195.17526550292968, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.027755821558336415, "epoch": 3.0483624119987756, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 510541219.0, "reward": 8.022947549819946, "reward_std": 0.3847667098045349, "rewards/combined_reward/mean": 8.022947645187378, "rewards/combined_reward/std": 2.2197011351585387, "step": 4980 }, { "clip_ratio/high_max": 0.002321438325937682, "clip_ratio/high_mean": 0.0012592258241056697, "clip_ratio/low_mean": 0.0008714326777408132, "clip_ratio/low_min": 0.00022340151820875084, "clip_ratio/region_mean": 0.0021306585035442064, "completions/clipped_ratio": 0.0, "completions/max_length": 523.3, "completions/max_terminated_length": 523.3, "completions/mean_length": 195.98985137939454, "completions/mean_terminated_length": 195.98985137939454, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.026868691861939926, "epoch": 3.066727884909703, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 513593980.0, "reward": 8.069220066070557, "reward_std": 0.3316150948405266, "rewards/combined_reward/mean": 8.069219779968261, "rewards/combined_reward/std": 2.3458720684051513, "step": 5010 }, { "clip_ratio/high_max": 0.0022589224652620034, "clip_ratio/high_mean": 0.0012726185018739973, "clip_ratio/low_mean": 0.0009820252898255907, "clip_ratio/low_min": 0.00027054979752089516, "clip_ratio/region_mean": 0.002254643788425407, "completions/clipped_ratio": 0.0, "completions/max_length": 621.5, "completions/max_terminated_length": 621.5, "completions/mean_length": 210.03724365234376, "completions/mean_terminated_length": 210.03724365234376, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.027904467020804683, "epoch": 3.0850933578206305, "frac_reward_zero_std": 0.16666667126119136, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 516700683.0, "reward": 8.08648452758789, "reward_std": 0.3692895233631134, "rewards/combined_reward/mean": 8.086484289169311, "rewards/combined_reward/std": 2.1820563077926636, "step": 5040 }, { "clip_ratio/high_max": 0.0026887404926431674, "clip_ratio/high_mean": 0.001493867363751633, "clip_ratio/low_mean": 0.0008713584763124042, "clip_ratio/low_min": 0.00026201934088021517, "clip_ratio/region_mean": 0.002365225834849601, "completions/clipped_ratio": 0.0, "completions/max_length": 759.5, "completions/max_terminated_length": 759.5, "completions/mean_length": 210.35677642822264, "completions/mean_terminated_length": 210.35677642822264, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.031967558059841396, "epoch": 3.103458830731558, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 519808613.0, "reward": 8.099337959289551, "reward_std": 0.36987523436546327, "rewards/combined_reward/mean": 8.09933786392212, "rewards/combined_reward/std": 2.3676317811012266, "step": 5070 }, { "clip_ratio/high_max": 0.0025386274738896947, "clip_ratio/high_mean": 0.0013946334773208946, "clip_ratio/low_mean": 0.0009865668732042347, "clip_ratio/low_min": 0.00029765297998286165, "clip_ratio/region_mean": 0.00238120034749348, "completions/clipped_ratio": 0.0, "completions/max_length": 622.5, "completions/max_terminated_length": 622.5, "completions/mean_length": 209.30156860351562, "completions/mean_terminated_length": 209.30156860351562, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.030893510937069855, "epoch": 3.1218243036424855, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 522912491.0, "reward": 8.159891653060914, "reward_std": 0.3653822332620621, "rewards/combined_reward/mean": 8.159891271591187, "rewards/combined_reward/std": 2.1358582258224486, "step": 5100 }, { "clip_ratio/high_max": 0.002299098666117061, "clip_ratio/high_mean": 0.0013367669918807224, "clip_ratio/low_mean": 0.0010216891364204154, "clip_ratio/low_min": 0.0003376908025529701, "clip_ratio/region_mean": 0.0023584561349707656, "completions/clipped_ratio": 0.0, "completions/max_length": 566.9, "completions/max_terminated_length": 566.9, "completions/mean_length": 208.43359832763673, "completions/mean_terminated_length": 208.43359832763673, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02997582551712791, "epoch": 3.140189776553413, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.03076171875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 526013036.0, "reward": 8.051339244842529, "reward_std": 0.37989163100719453, "rewards/combined_reward/mean": 8.051339054107666, "rewards/combined_reward/std": 2.346064829826355, "step": 5130 }, { "clip_ratio/high_max": 0.002314718322789607, "clip_ratio/high_mean": 0.001307108894494983, "clip_ratio/low_mean": 0.0011009929950887454, "clip_ratio/low_min": 0.0003651216963286667, "clip_ratio/region_mean": 0.002408101896192723, "completions/clipped_ratio": 0.0, "completions/max_length": 568.5, "completions/max_terminated_length": 568.5, "completions/mean_length": 201.18984832763672, "completions/mean_terminated_length": 201.18984832763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.032013976961995164, "epoch": 3.1585552494643405, "frac_reward_zero_std": 0.13333333656191826, "grad_norm": 0.005218505859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 529085765.0, "reward": 8.142257833480835, "reward_std": 0.32482413947582245, "rewards/combined_reward/mean": 8.142257690429688, "rewards/combined_reward/std": 2.033657598495483, "step": 5160 }, { "clip_ratio/high_max": 0.002104254433652386, "clip_ratio/high_mean": 0.00119177998288554, "clip_ratio/low_mean": 0.0008977340860534847, "clip_ratio/low_min": 0.0002632733216159977, "clip_ratio/region_mean": 0.0020895140837334718, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 210.80365142822265, "completions/mean_terminated_length": 210.80365142822265, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.026341072873522837, "epoch": 3.176920722375268, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 532195411.0, "reward": 8.242109203338623, "reward_std": 0.32477460652589796, "rewards/combined_reward/mean": 8.242109155654907, "rewards/combined_reward/std": 2.109383535385132, "step": 5190 }, { "clip_ratio/high_max": 0.002616349693077306, "clip_ratio/high_mean": 0.0015104147908762873, "clip_ratio/low_mean": 0.0010053531508674497, "clip_ratio/low_min": 0.0003192610148592697, "clip_ratio/region_mean": 0.0025157679182787737, "completions/clipped_ratio": 0.0, "completions/max_length": 634.7, "completions/max_terminated_length": 634.7, "completions/mean_length": 190.3726593017578, "completions/mean_terminated_length": 190.3726593017578, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030349081025148432, "epoch": 3.1952861952861955, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.006072998046875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 535226602.0, "reward": 8.113804578781128, "reward_std": 0.31306917518377303, "rewards/combined_reward/mean": 8.113804149627686, "rewards/combined_reward/std": 2.2881505608558657, "step": 5220 }, { "clip_ratio/high_max": 0.002228604684448025, "clip_ratio/high_mean": 0.0012846246793439301, "clip_ratio/low_mean": 0.0008834724010133262, "clip_ratio/low_min": 0.00019419656479537176, "clip_ratio/region_mean": 0.0021680970831463734, "completions/clipped_ratio": 0.0, "completions/max_length": 583.9, "completions/max_terminated_length": 583.9, "completions/mean_length": 197.29662170410157, "completions/mean_terminated_length": 197.29662170410157, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.025899839028716086, "epoch": 3.2136516681971226, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 538284381.0, "reward": 8.131101369857788, "reward_std": 0.32468149065971375, "rewards/combined_reward/mean": 8.131100988388061, "rewards/combined_reward/std": 2.189035654067993, "step": 5250 }, { "clip_ratio/high_max": 0.0022487934572078907, "clip_ratio/high_mean": 0.0012957128389340748, "clip_ratio/low_mean": 0.0009802017094746893, "clip_ratio/low_min": 0.0001485613024366709, "clip_ratio/region_mean": 0.0022759145625362485, "completions/clipped_ratio": 0.0, "completions/max_length": 697.5, "completions/max_terminated_length": 697.5, "completions/mean_length": 204.7502685546875, "completions/mean_terminated_length": 204.7502685546875, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.02799598821438849, "epoch": 3.23201714110805, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 541370782.0, "reward": 7.9808708190917965, "reward_std": 0.4602570176124573, "rewards/combined_reward/mean": 7.980870628356934, "rewards/combined_reward/std": 2.393754744529724, "step": 5280 }, { "clip_ratio/high_max": 0.002194094575437096, "clip_ratio/high_mean": 0.0012569982860441086, "clip_ratio/low_mean": 0.0009491232266251852, "clip_ratio/low_min": 0.00021515443028571706, "clip_ratio/region_mean": 0.002206121506363464, "completions/clipped_ratio": 0.0, "completions/max_length": 532.7, "completions/max_terminated_length": 532.7, "completions/mean_length": 194.79505615234376, "completions/mean_terminated_length": 194.79505615234376, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.02874210267327726, "epoch": 3.2503826140189775, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0076904296875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 544418955.0, "reward": 7.999461984634399, "reward_std": 0.3209122568368912, "rewards/combined_reward/mean": 7.9994617938995365, "rewards/combined_reward/std": 2.274247181415558, "step": 5310 }, { "clip_ratio/high_max": 0.0027640237160085233, "clip_ratio/high_mean": 0.0015749531938733222, "clip_ratio/low_mean": 0.0009453421904860685, "clip_ratio/low_min": 0.00023492822559395183, "clip_ratio/region_mean": 0.0025202953882399017, "completions/clipped_ratio": 0.0, "completions/max_length": 583.6, "completions/max_terminated_length": 583.6, "completions/mean_length": 209.10677642822264, "completions/mean_terminated_length": 209.10677642822264, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.03312639365904033, "epoch": 3.268748086929905, "frac_reward_zero_std": 0.1375000048428774, "grad_norm": 0.0223388671875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 547522085.0, "reward": 8.24277491569519, "reward_std": 0.419348406791687, "rewards/combined_reward/mean": 8.242774820327758, "rewards/combined_reward/std": 1.9957989573478698, "step": 5340 }, { "clip_ratio/high_max": 0.0023884492329671048, "clip_ratio/high_mean": 0.001317394388752291, "clip_ratio/low_mean": 0.0007942106517778787, "clip_ratio/low_min": 0.0002153744285654587, "clip_ratio/region_mean": 0.0021116050285248395, "completions/clipped_ratio": 0.0, "completions/max_length": 561.1, "completions/max_terminated_length": 561.1, "completions/mean_length": 191.6893280029297, "completions/mean_terminated_length": 191.6893280029297, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.02814455480935673, "epoch": 3.2871135598408325, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.04248046875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 550558332.0, "reward": 7.8310645580291744, "reward_std": 0.4151975572109222, "rewards/combined_reward/mean": 7.831064224243164, "rewards/combined_reward/std": 2.472068691253662, "step": 5370 }, { "clip_ratio/high_max": 0.0029664547090457443, "clip_ratio/high_mean": 0.0017111118709484193, "clip_ratio/low_mean": 0.0009730503755539152, "clip_ratio/low_min": 0.00026818387656627846, "clip_ratio/region_mean": 0.0026841622340725733, "completions/clipped_ratio": 0.0, "completions/max_length": 601.4, "completions/max_terminated_length": 601.4, "completions/mean_length": 196.3000045776367, "completions/mean_terminated_length": 196.3000045776367, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.03339463984593749, "epoch": 3.30547903275176, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.0042724609375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 553612284.0, "reward": 8.139923572540283, "reward_std": 0.35465587228536605, "rewards/combined_reward/mean": 8.139923477172852, "rewards/combined_reward/std": 2.275649404525757, "step": 5400 }, { "clip_ratio/high_max": 0.0025168837513774634, "clip_ratio/high_mean": 0.0014912285027094186, "clip_ratio/low_mean": 0.001023848610323815, "clip_ratio/low_min": 0.00022984260673789927, "clip_ratio/region_mean": 0.0025150770952071374, "completions/clipped_ratio": 0.0, "completions/max_length": 642.7, "completions/max_terminated_length": 642.7, "completions/mean_length": 200.4484436035156, "completions/mean_terminated_length": 200.4484436035156, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.03489424486955007, "epoch": 3.3238445056626875, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 556682166.0, "reward": 7.934696340560913, "reward_std": 0.2996515825390816, "rewards/combined_reward/mean": 7.934696054458618, "rewards/combined_reward/std": 2.375451755523682, "step": 5430 }, { "clip_ratio/high_max": 0.002470184176248343, "clip_ratio/high_mean": 0.001428827363270102, "clip_ratio/low_mean": 0.0009466227769735269, "clip_ratio/low_min": 0.0002275945014844183, "clip_ratio/region_mean": 0.0023754501511575655, "completions/clipped_ratio": 0.0, "completions/max_length": 601.5, "completions/max_terminated_length": 601.5, "completions/mean_length": 190.88125457763672, "completions/mean_terminated_length": 190.88125457763672, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.028761676947275796, "epoch": 3.342209978573615, "frac_reward_zero_std": 0.1708333406597376, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 559715310.0, "reward": 8.258774900436402, "reward_std": 0.27708870768547056, "rewards/combined_reward/mean": 8.258774757385254, "rewards/combined_reward/std": 2.0299688935279847, "step": 5460 }, { "clip_ratio/high_max": 0.002467339028468511, "clip_ratio/high_mean": 0.0014009148433009007, "clip_ratio/low_mean": 0.0009619224788669574, "clip_ratio/low_min": 0.0003042295723086378, "clip_ratio/region_mean": 0.0023628373218040604, "completions/clipped_ratio": 0.0, "completions/max_length": 637.6, "completions/max_terminated_length": 637.6, "completions/mean_length": 203.80599365234374, "completions/mean_terminated_length": 203.80599365234374, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0284752421391507, "epoch": 3.3605754514845425, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 562798085.0, "reward": 8.196513748168945, "reward_std": 0.34671002849936483, "rewards/combined_reward/mean": 8.196513462066651, "rewards/combined_reward/std": 2.0625844836235045, "step": 5490 }, { "clip_ratio/high_max": 0.0024349484864311915, "clip_ratio/high_mean": 0.0013903526201223333, "clip_ratio/low_mean": 0.0010003672359744087, "clip_ratio/low_min": 0.000293882998812478, "clip_ratio/region_mean": 0.002390719853186359, "completions/clipped_ratio": 0.0, "completions/max_length": 523.7, "completions/max_terminated_length": 523.7, "completions/mean_length": 197.23020935058594, "completions/mean_terminated_length": 197.23020935058594, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.030348435960089168, "epoch": 3.37894092439547, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 565855609.0, "reward": 8.193823337554932, "reward_std": 0.32499777227640153, "rewards/combined_reward/mean": 8.193822956085205, "rewards/combined_reward/std": 2.230313503742218, "step": 5520 }, { "clip_ratio/high_max": 0.002159653430377754, "clip_ratio/high_mean": 0.0011662074825532424, "clip_ratio/low_mean": 0.0009090998452544833, "clip_ratio/low_min": 0.0001685787872702349, "clip_ratio/region_mean": 0.002075307337023939, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 191.54037017822264, "completions/mean_terminated_length": 191.54037017822264, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.026708185905590653, "epoch": 3.3973063973063975, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.004791259765625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 568891284.0, "reward": 8.357570934295655, "reward_std": 0.23496017903089522, "rewards/combined_reward/mean": 8.357570457458497, "rewards/combined_reward/std": 1.8740468144416809, "step": 5550 }, { "clip_ratio/high_max": 0.0020611704142841822, "clip_ratio/high_mean": 0.0011358271968977837, "clip_ratio/low_mean": 0.0008983540820433215, "clip_ratio/low_min": 0.0001865920625277795, "clip_ratio/region_mean": 0.002034181274454265, "completions/clipped_ratio": 0.0, "completions/max_length": 691.6, "completions/max_terminated_length": 691.6, "completions/mean_length": 195.7085983276367, "completions/mean_terminated_length": 195.7085983276367, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.025820783580032487, "epoch": 3.4156718702173245, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 571942965.0, "reward": 7.931357526779175, "reward_std": 0.3166040673851967, "rewards/combined_reward/mean": 7.931357383728027, "rewards/combined_reward/std": 2.2883009791374205, "step": 5580 }, { "clip_ratio/high_max": 0.0027264218724061115, "clip_ratio/high_mean": 0.001528754560301119, "clip_ratio/low_mean": 0.0008773999402061842, "clip_ratio/low_min": 0.000186775193772822, "clip_ratio/region_mean": 0.002406154510633011, "completions/clipped_ratio": 0.0, "completions/max_length": 544.4, "completions/max_terminated_length": 544.4, "completions/mean_length": 200.40234985351563, "completions/mean_terminated_length": 200.40234985351563, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.0298045920751368, "epoch": 3.434037343128252, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 575012670.0, "reward": 8.07356653213501, "reward_std": 0.3686544582247734, "rewards/combined_reward/mean": 8.073566436767578, "rewards/combined_reward/std": 2.4171477794647216, "step": 5610 }, { "clip_ratio/high_max": 0.0021291105683000446, "clip_ratio/high_mean": 0.0012126441829119964, "clip_ratio/low_mean": 0.0009832237129254887, "clip_ratio/low_min": 0.0002719423903423982, "clip_ratio/region_mean": 0.002195867888318996, "completions/clipped_ratio": 0.0, "completions/max_length": 572.7, "completions/max_terminated_length": 572.7, "completions/mean_length": 192.84062805175782, "completions/mean_terminated_length": 192.84062805175782, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.02875271754649778, "epoch": 3.4524028160391795, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.025390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 578053338.0, "reward": 7.844840812683105, "reward_std": 0.33809482902288435, "rewards/combined_reward/mean": 7.844840621948242, "rewards/combined_reward/std": 2.413142669200897, "step": 5640 }, { "clip_ratio/high_max": 0.00250756214033269, "clip_ratio/high_mean": 0.0013666432321770116, "clip_ratio/low_mean": 0.0007748612514357471, "clip_ratio/low_min": 0.00020583123744775852, "clip_ratio/region_mean": 0.002141504501923919, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 204.64792175292968, "completions/mean_terminated_length": 204.64792175292968, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.026692416581014792, "epoch": 3.470768288950107, "frac_reward_zero_std": 0.12916667126119136, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 581139346.0, "reward": 8.162485504150391, "reward_std": 0.32477535158395765, "rewards/combined_reward/mean": 8.162485074996948, "rewards/combined_reward/std": 2.3036073207855225, "step": 5670 }, { "clip_ratio/high_max": 0.0024436967476503925, "clip_ratio/high_mean": 0.0013043084922780205, "clip_ratio/low_mean": 0.000984239935739121, "clip_ratio/low_min": 0.00024431087634487386, "clip_ratio/region_mean": 0.0022885484586974295, "completions/clipped_ratio": 0.0, "completions/max_length": 623.9, "completions/max_terminated_length": 623.9, "completions/mean_length": 188.2612045288086, "completions/mean_terminated_length": 188.2612045288086, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.029471345975374184, "epoch": 3.4891337618610345, "frac_reward_zero_std": 0.1750000063329935, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 584162429.0, "reward": 8.019039821624755, "reward_std": 0.33535448759794234, "rewards/combined_reward/mean": 8.019039344787597, "rewards/combined_reward/std": 2.4323641061782837, "step": 5700 }, { "clip_ratio/high_max": 0.0027450023762260875, "clip_ratio/high_mean": 0.0014627386789167454, "clip_ratio/low_mean": 0.0009221382204486872, "clip_ratio/low_min": 0.0002684679398953449, "clip_ratio/region_mean": 0.0023848768993047996, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 194.95052490234374, "completions/mean_terminated_length": 194.95052490234374, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.030227117504303656, "epoch": 3.507499234771962, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 587211199.0, "reward": 8.039965438842774, "reward_std": 0.35978697538375853, "rewards/combined_reward/mean": 8.039965152740479, "rewards/combined_reward/std": 2.27087162733078, "step": 5730 }, { "clip_ratio/high_max": 0.0028721766391148168, "clip_ratio/high_mean": 0.001639382034530475, "clip_ratio/low_mean": 0.0011717627117225978, "clip_ratio/low_min": 0.00032763602794148026, "clip_ratio/region_mean": 0.0028111447347328065, "completions/clipped_ratio": 0.0, "completions/max_length": 588.3, "completions/max_terminated_length": 588.3, "completions/mean_length": 198.82109680175782, "completions/mean_terminated_length": 198.82109680175782, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.037705186769987145, "epoch": 3.5258647076828895, "frac_reward_zero_std": 0.20000000558793546, "grad_norm": 0.004058837890625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 590274832.0, "reward": 7.887068605422973, "reward_std": 0.2914671190083027, "rewards/combined_reward/mean": 7.887068605422973, "rewards/combined_reward/std": 2.5043334007263183, "step": 5760 }, { "clip_ratio/high_max": 0.0020249907038911866, "clip_ratio/high_mean": 0.001183188264258206, "clip_ratio/low_mean": 0.0008880944348372092, "clip_ratio/low_min": 0.0003081550591256625, "clip_ratio/region_mean": 0.00207128270509808, "completions/clipped_ratio": 0.0, "completions/max_length": 639.4, "completions/max_terminated_length": 639.4, "completions/mean_length": 206.9523483276367, "completions/mean_terminated_length": 206.9523483276367, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.02703253731597215, "epoch": 3.544230180593817, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 593369689.0, "reward": 8.375552320480347, "reward_std": 0.3533032774925232, "rewards/combined_reward/mean": 8.375552225112916, "rewards/combined_reward/std": 1.8991389155387879, "step": 5790 }, { "clip_ratio/high_max": 0.002622041154730444, "clip_ratio/high_mean": 0.0015569702673625821, "clip_ratio/low_mean": 0.000998657589176825, "clip_ratio/low_min": 0.00026610834611346946, "clip_ratio/region_mean": 0.002555627841502428, "completions/clipped_ratio": 0.0, "completions/max_length": 593.8, "completions/max_terminated_length": 593.8, "completions/mean_length": 200.39688262939453, "completions/mean_terminated_length": 200.39688262939453, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.029919973574578762, "epoch": 3.5625956535047445, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0107421875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 596439373.0, "reward": 7.990189456939698, "reward_std": 0.3357082918286324, "rewards/combined_reward/mean": 7.990189123153686, "rewards/combined_reward/std": 2.207187533378601, "step": 5820 }, { "clip_ratio/high_max": 0.002742065759957768, "clip_ratio/high_mean": 0.0015219730286238095, "clip_ratio/low_mean": 0.000939219756037346, "clip_ratio/low_min": 0.00028178868087707086, "clip_ratio/region_mean": 0.002461192785995081, "completions/clipped_ratio": 0.0, "completions/max_length": 587.7, "completions/max_terminated_length": 587.7, "completions/mean_length": 202.08958892822267, "completions/mean_terminated_length": 202.08958892822267, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.030970921708891787, "epoch": 3.580961126415672, "frac_reward_zero_std": 0.18750000782310963, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 599515557.0, "reward": 8.033844423294067, "reward_std": 0.34530273675918577, "rewards/combined_reward/mean": 8.033844232559204, "rewards/combined_reward/std": 2.182740330696106, "step": 5850 }, { "clip_ratio/high_max": 0.002119657750396679, "clip_ratio/high_mean": 0.0012445619013305986, "clip_ratio/low_mean": 0.0008117984862716791, "clip_ratio/low_min": 0.00022565814603391725, "clip_ratio/region_mean": 0.0020563603757182136, "completions/clipped_ratio": 0.0, "completions/max_length": 621.4, "completions/max_terminated_length": 621.4, "completions/mean_length": 204.37500457763673, "completions/mean_terminated_length": 204.37500457763673, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.026501512822384634, "epoch": 3.5993265993265995, "frac_reward_zero_std": 0.1750000063329935, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 602600517.0, "reward": 8.456466627120971, "reward_std": 0.31904478222131727, "rewards/combined_reward/mean": 8.456466245651246, "rewards/combined_reward/std": 1.898151409626007, "step": 5880 }, { "clip_ratio/high_max": 0.002497296595538501, "clip_ratio/high_mean": 0.0013909608222699413, "clip_ratio/low_mean": 0.0007779856758133973, "clip_ratio/low_min": 0.00016581757469490792, "clip_ratio/region_mean": 0.002168946510452467, "completions/clipped_ratio": 0.0, "completions/max_length": 554.5, "completions/max_terminated_length": 554.5, "completions/mean_length": 194.4674545288086, "completions/mean_terminated_length": 194.4674545288086, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.02717176922596991, "epoch": 3.617692072237527, "frac_reward_zero_std": 0.15833333991467952, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 605647432.0, "reward": 7.963862419128418, "reward_std": 0.3501826286315918, "rewards/combined_reward/mean": 7.963862085342408, "rewards/combined_reward/std": 2.3259896874427795, "step": 5910 }, { "clip_ratio/high_max": 0.002862122648123962, "clip_ratio/high_mean": 0.0014983589554807017, "clip_ratio/low_mean": 0.0009568900258576225, "clip_ratio/low_min": 0.000265473500379206, "clip_ratio/region_mean": 0.0024552489863708614, "completions/clipped_ratio": 0.0, "completions/max_length": 550.7, "completions/max_terminated_length": 550.7, "completions/mean_length": 200.08307800292968, "completions/mean_terminated_length": 200.08307800292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.029051849137370786, "epoch": 3.6360575451484545, "frac_reward_zero_std": 0.18750000558793545, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 608715911.0, "reward": 7.932967042922973, "reward_std": 0.3379728376865387, "rewards/combined_reward/mean": 7.9329667568206785, "rewards/combined_reward/std": 2.2237315654754637, "step": 5940 }, { "clip_ratio/high_max": 0.0021237443696008997, "clip_ratio/high_mean": 0.0012337892721310103, "clip_ratio/low_mean": 0.0009764689144503791, "clip_ratio/low_min": 0.00030493624605393657, "clip_ratio/region_mean": 0.0022102581919170917, "completions/clipped_ratio": 0.0, "completions/max_length": 566.6, "completions/max_terminated_length": 566.6, "completions/mean_length": 200.64844512939453, "completions/mean_terminated_length": 200.64844512939453, "completions/min_length": 45.6, "completions/min_terminated_length": 45.6, "entropy": 0.028462106866451602, "epoch": 3.654423018059382, "frac_reward_zero_std": 0.12500000447034837, "grad_norm": 0.0047607421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 611786561.0, "reward": 7.982969331741333, "reward_std": 0.31417706310749055, "rewards/combined_reward/mean": 7.9829692363739015, "rewards/combined_reward/std": 2.2720341444015504, "step": 5970 }, { "clip_ratio/high_max": 0.002607413813530002, "clip_ratio/high_mean": 0.0014726377456099726, "clip_ratio/low_mean": 0.0009480811281415906, "clip_ratio/low_min": 0.0002169551439389276, "clip_ratio/region_mean": 0.002420718865323579, "completions/clipped_ratio": 0.0, "completions/max_length": 602.8, "completions/max_terminated_length": 602.8, "completions/mean_length": 196.16641235351562, "completions/mean_terminated_length": 196.16641235351562, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.028135747959216435, "epoch": 3.672788490970309, "frac_reward_zero_std": 0.19166667051613331, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 614840000.0, "reward": 8.068321418762206, "reward_std": 0.2901913642883301, "rewards/combined_reward/mean": 8.068321371078492, "rewards/combined_reward/std": 2.247336673736572, "step": 6000 }, { "clip_ratio/high_max": 0.0024224239226896317, "clip_ratio/high_mean": 0.0014344930517836475, "clip_ratio/low_mean": 0.0010348493780232579, "clip_ratio/low_min": 0.0003106777818175033, "clip_ratio/region_mean": 0.002469342435748937, "completions/clipped_ratio": 0.0, "completions/max_length": 668.5, "completions/max_terminated_length": 668.5, "completions/mean_length": 209.49714202880858, "completions/mean_terminated_length": 209.49714202880858, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029499471358334025, "epoch": 3.6911539638812365, "frac_reward_zero_std": 0.12083333618938923, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 617944629.0, "reward": 7.884818077087402, "reward_std": 0.48571284413337706, "rewards/combined_reward/mean": 7.884817886352539, "rewards/combined_reward/std": 2.5527934789657594, "step": 6030 }, { "clip_ratio/high_max": 0.002707660676484617, "clip_ratio/high_mean": 0.0015277088663424366, "clip_ratio/low_mean": 0.0010833242128380032, "clip_ratio/low_min": 0.00029306044404317315, "clip_ratio/region_mean": 0.0026110330858500675, "completions/clipped_ratio": 0.0, "completions/max_length": 714.3, "completions/max_terminated_length": 714.3, "completions/mean_length": 205.6682327270508, "completions/mean_terminated_length": 205.6682327270508, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.03231038280452291, "epoch": 3.709519436792164, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 621034555.0, "reward": 7.901114511489868, "reward_std": 0.35342108458280563, "rewards/combined_reward/mean": 7.90111403465271, "rewards/combined_reward/std": 2.391073477268219, "step": 6060 }, { "clip_ratio/high_max": 0.002302383282221854, "clip_ratio/high_mean": 0.001271786964207422, "clip_ratio/low_mean": 0.0010095771591295487, "clip_ratio/low_min": 0.00028499208516829334, "clip_ratio/region_mean": 0.0022813641374038224, "completions/clipped_ratio": 0.0, "completions/max_length": 623.7, "completions/max_terminated_length": 623.7, "completions/mean_length": 207.61380767822266, "completions/mean_terminated_length": 207.61380767822266, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03018238744698465, "epoch": 3.7278849097030915, "frac_reward_zero_std": 0.13333333432674407, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 624131952.0, "reward": 7.953459978103638, "reward_std": 0.3831681877374649, "rewards/combined_reward/mean": 7.953459739685059, "rewards/combined_reward/std": 2.5810283064842223, "step": 6090 }, { "clip_ratio/high_max": 0.002355090741669604, "clip_ratio/high_mean": 0.001295326386753004, "clip_ratio/low_mean": 0.0010271702451670232, "clip_ratio/low_min": 0.0003040618025503742, "clip_ratio/region_mean": 0.002322496644531687, "completions/clipped_ratio": 0.0, "completions/max_length": 552.4, "completions/max_terminated_length": 552.4, "completions/mean_length": 206.7182357788086, "completions/mean_terminated_length": 206.7182357788086, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028666269298021994, "epoch": 3.746250382614019, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 627225910.0, "reward": 7.976446723937988, "reward_std": 0.37551163882017136, "rewards/combined_reward/mean": 7.976446580886841, "rewards/combined_reward/std": 2.3090658664703367, "step": 6120 }, { "clip_ratio/high_max": 0.002804985324716351, "clip_ratio/high_mean": 0.0017355715674057138, "clip_ratio/low_mean": 0.0011829966423344254, "clip_ratio/low_min": 0.0003518210660937863, "clip_ratio/region_mean": 0.002918568213256852, "completions/clipped_ratio": 0.0, "completions/max_length": 602.8, "completions/max_terminated_length": 602.8, "completions/mean_length": 210.17344360351564, "completions/mean_terminated_length": 210.17344360351564, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.03557930279833575, "epoch": 3.7646158555249465, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0299072265625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 630333136.0, "reward": 7.733653926849366, "reward_std": 0.40994020700454714, "rewards/combined_reward/mean": 7.73365364074707, "rewards/combined_reward/std": 2.476205599308014, "step": 6150 }, { "clip_ratio/high_max": 0.0025294717221792475, "clip_ratio/high_mean": 0.0015031486237906697, "clip_ratio/low_mean": 0.0011640610477115843, "clip_ratio/low_min": 0.000422055140613035, "clip_ratio/region_mean": 0.002667209690359111, "completions/clipped_ratio": 0.0, "completions/max_length": 706.9, "completions/max_terminated_length": 706.9, "completions/mean_length": 204.60261077880858, "completions/mean_terminated_length": 204.60261077880858, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03390380904699365, "epoch": 3.782981328435874, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 633418970.0, "reward": 7.935478019714355, "reward_std": 0.368173149228096, "rewards/combined_reward/mean": 7.935477590560913, "rewards/combined_reward/std": 2.3742892384529113, "step": 6180 }, { "clip_ratio/high_max": 0.0024736718444425304, "clip_ratio/high_mean": 0.001490661945232811, "clip_ratio/low_mean": 0.0011654577632725705, "clip_ratio/low_min": 0.0003513480473581391, "clip_ratio/region_mean": 0.0026561197106881687, "completions/clipped_ratio": 0.0, "completions/max_length": 604.5, "completions/max_terminated_length": 604.5, "completions/mean_length": 194.5276077270508, "completions/mean_terminated_length": 194.5276077270508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.031068458252896867, "epoch": 3.8013468013468015, "frac_reward_zero_std": 0.22083333730697632, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 636466116.0, "reward": 7.829420328140259, "reward_std": 0.3618571162223816, "rewards/combined_reward/mean": 7.829420280456543, "rewards/combined_reward/std": 2.3704976797103883, "step": 6210 }, { "clip_ratio/high_max": 0.0026900913551799023, "clip_ratio/high_mean": 0.0014913419433772409, "clip_ratio/low_mean": 0.0011239448548925187, "clip_ratio/low_min": 0.00035971304023405535, "clip_ratio/region_mean": 0.002615286808577366, "completions/clipped_ratio": 0.0, "completions/max_length": 501.4, "completions/max_terminated_length": 501.4, "completions/mean_length": 200.62318267822266, "completions/mean_terminated_length": 200.62318267822266, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.032277327543124554, "epoch": 3.819712274257729, "frac_reward_zero_std": 0.1500000063329935, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 639536669.0, "reward": 8.175046491622926, "reward_std": 0.3262510061264038, "rewards/combined_reward/mean": 8.175046443939209, "rewards/combined_reward/std": 2.0678364515304564, "step": 6240 }, { "clip_ratio/high_max": 0.00258500373311108, "clip_ratio/high_mean": 0.0015950137191490892, "clip_ratio/low_mean": 0.0011843764411120599, "clip_ratio/low_min": 0.0003216259356122464, "clip_ratio/region_mean": 0.0027793901788148406, "completions/clipped_ratio": 0.0, "completions/max_length": 613.4, "completions/max_terminated_length": 613.4, "completions/mean_length": 205.04063110351564, "completions/mean_terminated_length": 205.04063110351564, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03152101336357494, "epoch": 3.838077747168656, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.0283203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 642624185.0, "reward": 7.555079698562622, "reward_std": 0.49450784623622895, "rewards/combined_reward/mean": 7.555079507827759, "rewards/combined_reward/std": 2.638063979148865, "step": 6270 }, { "clip_ratio/high_max": 0.0026795870755449867, "clip_ratio/high_mean": 0.0014838049411385632, "clip_ratio/low_mean": 0.0010859169718969498, "clip_ratio/low_min": 0.00031530489650322124, "clip_ratio/region_mean": 0.002569721927284263, "completions/clipped_ratio": 0.0, "completions/max_length": 656.1, "completions/max_terminated_length": 656.1, "completions/mean_length": 201.25443115234376, "completions/mean_terminated_length": 201.25443115234376, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03220579079352319, "epoch": 3.8564432200795835, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 645697162.0, "reward": 8.084075832366944, "reward_std": 0.37564343214035034, "rewards/combined_reward/mean": 8.084075546264648, "rewards/combined_reward/std": 2.182408392429352, "step": 6300 }, { "clip_ratio/high_max": 0.002367574894257511, "clip_ratio/high_mean": 0.0013502161426004022, "clip_ratio/low_mean": 0.0011679145273471173, "clip_ratio/low_min": 0.000344893423365041, "clip_ratio/region_mean": 0.002518130650181168, "completions/clipped_ratio": 0.0, "completions/max_length": 555.9, "completions/max_terminated_length": 555.9, "completions/mean_length": 193.3276107788086, "completions/mean_terminated_length": 193.3276107788086, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.03060542034606139, "epoch": 3.874808692990511, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 648739700.0, "reward": 7.91180419921875, "reward_std": 0.38463967889547346, "rewards/combined_reward/mean": 7.9118040084838865, "rewards/combined_reward/std": 2.371174430847168, "step": 6330 }, { "clip_ratio/high_max": 0.002405529963046623, "clip_ratio/high_mean": 0.0014076047863151567, "clip_ratio/low_mean": 0.0011732857452443567, "clip_ratio/low_min": 0.00036808593261715336, "clip_ratio/region_mean": 0.002580890514461013, "completions/clipped_ratio": 0.0, "completions/max_length": 620.5, "completions/max_terminated_length": 620.5, "completions/mean_length": 198.87891235351563, "completions/mean_terminated_length": 198.87891235351563, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02898910144964854, "epoch": 3.8931741659014385, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 651803555.0, "reward": 7.621174097061157, "reward_std": 0.38936863988637926, "rewards/combined_reward/mean": 7.621173954010009, "rewards/combined_reward/std": 2.4840185761451723, "step": 6360 }, { "clip_ratio/high_max": 0.0023444237590107757, "clip_ratio/high_mean": 0.001292517653928371, "clip_ratio/low_mean": 0.0010800302796269534, "clip_ratio/low_min": 0.0003342272747734872, "clip_ratio/region_mean": 0.002372547935374314, "completions/clipped_ratio": 0.0, "completions/max_length": 559.4, "completions/max_terminated_length": 559.4, "completions/mean_length": 206.58932800292968, "completions/mean_terminated_length": 206.58932800292968, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.029837682261131704, "epoch": 3.911539638812366, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 654897018.0, "reward": 7.9760607242584225, "reward_std": 0.36221654117107394, "rewards/combined_reward/mean": 7.976060390472412, "rewards/combined_reward/std": 2.4018463015556337, "step": 6390 }, { "clip_ratio/high_max": 0.0020171104319160805, "clip_ratio/high_mean": 0.0011387073922378478, "clip_ratio/low_mean": 0.0008064333425863879, "clip_ratio/low_min": 0.00021098309734952637, "clip_ratio/region_mean": 0.001945140733732842, "completions/clipped_ratio": 0.0, "completions/max_length": 558.4285714285714, "completions/max_terminated_length": 558.4285714285714, "completions/mean_length": 202.15848868233817, "completions/mean_terminated_length": 202.15848868233817, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.028377513843588532, "epoch": 3.929292929292929, "frac_reward_zero_std": 0.16666667003716742, "grad_norm": 0.00347900390625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 657976119.0, "reward": 8.093403884342738, "reward_std": 0.2839034391301019, "rewards/combined_reward/mean": 8.093403543744769, "rewards/combined_reward/std": 2.2293928180422102, "step": 6420 }, { "clip_ratio/high_max": 0.0017381672332703602, "clip_ratio/high_mean": 0.0009421202417191428, "clip_ratio/low_mean": 0.0007769418967654929, "clip_ratio/low_min": 0.0002393432701258765, "clip_ratio/region_mean": 0.0017190621447904656, "completions/clipped_ratio": 0.0, "completions/max_length": 548.2, "completions/max_terminated_length": 548.2, "completions/mean_length": 197.00156860351564, "completions/mean_terminated_length": 197.00156860351564, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.027580118908857306, "epoch": 3.9476584022038566, "frac_reward_zero_std": 0.20416667573153974, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 661032765.0, "reward": 8.056810426712037, "reward_std": 0.328906723856926, "rewards/combined_reward/mean": 8.05681037902832, "rewards/combined_reward/std": 2.077943539619446, "step": 6450 }, { "clip_ratio/high_max": 0.0018727101240074262, "clip_ratio/high_mean": 0.0010326940485053152, "clip_ratio/low_mean": 0.0007180856184883547, "clip_ratio/low_min": 0.00019053811411140487, "clip_ratio/region_mean": 0.0017507796719049415, "completions/clipped_ratio": 0.0, "completions/max_length": 553.1, "completions/max_terminated_length": 553.1, "completions/mean_length": 212.55755767822265, "completions/mean_terminated_length": 212.55755767822265, "completions/min_length": 48.8, "completions/min_terminated_length": 48.8, "entropy": 0.02736986520079275, "epoch": 3.966023875114784, "frac_reward_zero_std": 0.1416666705161333, "grad_norm": 0.006683349609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 664149146.0, "reward": 8.11883192062378, "reward_std": 0.29180571585893633, "rewards/combined_reward/mean": 8.118831443786622, "rewards/combined_reward/std": 2.095475935935974, "step": 6480 }, { "clip_ratio/high_max": 0.0019160927452806693, "clip_ratio/high_mean": 0.0010635486126072161, "clip_ratio/low_mean": 0.0008171119231216532, "clip_ratio/low_min": 0.00018411809069220907, "clip_ratio/region_mean": 0.0018806605041997197, "completions/clipped_ratio": 0.0, "completions/max_length": 647.3, "completions/max_terminated_length": 647.3, "completions/mean_length": 198.6737045288086, "completions/mean_terminated_length": 198.6737045288086, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.028648768551647663, "epoch": 3.9843893480257115, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 667212213.0, "reward": 8.217909812927246, "reward_std": 0.3315473347902298, "rewards/combined_reward/mean": 8.217909288406371, "rewards/combined_reward/std": 2.154792106151581, "step": 6510 }, { "clip_ratio/high_max": 0.0022686818589792265, "clip_ratio/high_mean": 0.0012179104533508328, "clip_ratio/low_mean": 0.0007612303452333435, "clip_ratio/low_min": 0.0002361168859709726, "clip_ratio/region_mean": 0.0019791408157602075, "completions/clipped_ratio": 0.0, "completions/max_length": 638.3, "completions/max_terminated_length": 638.3, "completions/mean_length": 209.93646545410155, "completions/mean_terminated_length": 209.93646545410155, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.029283146999898504, "epoch": 4.003060912151821, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 670318529.0, "reward": 7.900010871887207, "reward_std": 0.35847034156322477, "rewards/combined_reward/mean": 7.900010585784912, "rewards/combined_reward/std": 2.2822492122650146, "step": 6540 }, { "clip_ratio/high_max": 0.00262343895980545, "clip_ratio/high_mean": 0.0014579358551903472, "clip_ratio/low_mean": 0.0009719547602799139, "clip_ratio/low_min": 0.00018937548690397914, "clip_ratio/region_mean": 0.0024298906171073514, "completions/clipped_ratio": 0.0, "completions/max_length": 590.6, "completions/max_terminated_length": 590.6, "completions/mean_length": 199.43854522705078, "completions/mean_terminated_length": 199.43854522705078, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02720782277174294, "epoch": 4.021426385062749, "frac_reward_zero_std": 0.10833333618938923, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 673384533.0, "reward": 8.313000297546386, "reward_std": 0.3937214285135269, "rewards/combined_reward/mean": 8.313000059127807, "rewards/combined_reward/std": 2.0700897932052613, "step": 6570 }, { "clip_ratio/high_max": 0.0023260156602191275, "clip_ratio/high_mean": 0.0012504190606705378, "clip_ratio/low_mean": 0.0011046833513925472, "clip_ratio/low_min": 0.0003974964907683898, "clip_ratio/region_mean": 0.002355102412790681, "completions/clipped_ratio": 0.0, "completions/max_length": 630.4, "completions/max_terminated_length": 630.4, "completions/mean_length": 204.11953735351562, "completions/mean_terminated_length": 204.11953735351562, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.02860527322627604, "epoch": 4.039791857973676, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 676468512.0, "reward": 7.950871419906616, "reward_std": 0.36094825714826584, "rewards/combined_reward/mean": 7.950871276855469, "rewards/combined_reward/std": 2.3837578296661377, "step": 6600 }, { "clip_ratio/high_max": 0.002394229303172324, "clip_ratio/high_mean": 0.001296701572452245, "clip_ratio/low_mean": 0.0008782884943987786, "clip_ratio/low_min": 0.0003109940228265865, "clip_ratio/region_mean": 0.002174990065395832, "completions/clipped_ratio": 0.0, "completions/max_length": 517.3, "completions/max_terminated_length": 517.3, "completions/mean_length": 196.5421905517578, "completions/mean_terminated_length": 196.5421905517578, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.027600362369169792, "epoch": 4.058157330884604, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 679523394.0, "reward": 8.168661403656007, "reward_std": 0.3759860977530479, "rewards/combined_reward/mean": 8.168661069869994, "rewards/combined_reward/std": 2.076196885108948, "step": 6630 }, { "clip_ratio/high_max": 0.0021258191069743284, "clip_ratio/high_mean": 0.0012829041540195856, "clip_ratio/low_mean": 0.0008751681212743279, "clip_ratio/low_min": 0.00020537330031705398, "clip_ratio/region_mean": 0.0021580722896032967, "completions/clipped_ratio": 0.0, "completions/max_length": 569.3, "completions/max_terminated_length": 569.3, "completions/mean_length": 207.87838897705078, "completions/mean_terminated_length": 207.87838897705078, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029283834885184965, "epoch": 4.076522803795531, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 682621807.0, "reward": 8.163852739334107, "reward_std": 0.32842076420783994, "rewards/combined_reward/mean": 8.16385259628296, "rewards/combined_reward/std": 2.21025013923645, "step": 6660 }, { "clip_ratio/high_max": 0.002261101673260176, "clip_ratio/high_mean": 0.0013298558110060791, "clip_ratio/low_mean": 0.0008651686778951747, "clip_ratio/low_min": 0.00021186936767965867, "clip_ratio/region_mean": 0.0021950245058784883, "completions/clipped_ratio": 0.0, "completions/max_length": 623.6, "completions/max_terminated_length": 623.6, "completions/mean_length": 205.72318115234376, "completions/mean_terminated_length": 205.72318115234376, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028202945878729225, "epoch": 4.094888276706459, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 685711944.0, "reward": 8.056198644638062, "reward_std": 0.3678277149796486, "rewards/combined_reward/mean": 8.056198501586914, "rewards/combined_reward/std": 2.422277903556824, "step": 6690 }, { "clip_ratio/high_max": 0.002520814739788572, "clip_ratio/high_mean": 0.0014867394612034937, "clip_ratio/low_mean": 0.0009806987894990017, "clip_ratio/low_min": 0.0002973967505871163, "clip_ratio/region_mean": 0.0024674382507024954, "completions/clipped_ratio": 0.0, "completions/max_length": 656.6, "completions/max_terminated_length": 656.6, "completions/mean_length": 214.4596405029297, "completions/mean_terminated_length": 214.4596405029297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.034039479680359364, "epoch": 4.113253749617386, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 688835629.0, "reward": 8.222779512405396, "reward_std": 0.39607002437114713, "rewards/combined_reward/mean": 8.222778797149658, "rewards/combined_reward/std": 2.0489836931228638, "step": 6720 }, { "clip_ratio/high_max": 0.0021428833046229557, "clip_ratio/high_mean": 0.0012075413953425596, "clip_ratio/low_mean": 0.0010195717059104936, "clip_ratio/low_min": 0.0002938753917987924, "clip_ratio/region_mean": 0.0022271131116819256, "completions/clipped_ratio": 0.0, "completions/max_length": 636.8, "completions/max_terminated_length": 636.8, "completions/mean_length": 206.95807800292968, "completions/mean_terminated_length": 206.95807800292968, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.029985707175607484, "epoch": 4.131619222528314, "frac_reward_zero_std": 0.1708333406597376, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 691930508.0, "reward": 7.977191066741943, "reward_std": 0.30688305795192716, "rewards/combined_reward/mean": 7.977190780639648, "rewards/combined_reward/std": 2.377994883060455, "step": 6750 }, { "clip_ratio/high_max": 0.0025297677614920153, "clip_ratio/high_mean": 0.0014893504630890675, "clip_ratio/low_mean": 0.001000910085591992, "clip_ratio/low_min": 0.0003015922576499482, "clip_ratio/region_mean": 0.0024902605514701766, "completions/clipped_ratio": 0.0, "completions/max_length": 569.4, "completions/max_terminated_length": 569.4, "completions/mean_length": 199.96198425292968, "completions/mean_terminated_length": 199.96198425292968, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.03033532863482833, "epoch": 4.149984695439241, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 694998522.0, "reward": 8.165205955505371, "reward_std": 0.3199221551418304, "rewards/combined_reward/mean": 8.165205907821655, "rewards/combined_reward/std": 2.044101333618164, "step": 6780 }, { "clip_ratio/high_max": 0.0026872982591157777, "clip_ratio/high_mean": 0.0014224532960118571, "clip_ratio/low_mean": 0.000895178185783152, "clip_ratio/low_min": 0.0002154990354028996, "clip_ratio/region_mean": 0.0023176314774900676, "completions/clipped_ratio": 0.0, "completions/max_length": 572.8, "completions/max_terminated_length": 572.8, "completions/mean_length": 209.30990295410157, "completions/mean_terminated_length": 209.30990295410157, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029563402120644847, "epoch": 4.168350168350169, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 698102432.0, "reward": 8.293191480636597, "reward_std": 0.32815817147493365, "rewards/combined_reward/mean": 8.293190908432006, "rewards/combined_reward/std": 2.0188760280609133, "step": 6810 }, { "clip_ratio/high_max": 0.002377855919864184, "clip_ratio/high_mean": 0.0013497049487341429, "clip_ratio/low_mean": 0.001021728478614629, "clip_ratio/low_min": 0.00023535005263208101, "clip_ratio/region_mean": 0.00237143342771257, "completions/clipped_ratio": 0.0, "completions/max_length": 617.8, "completions/max_terminated_length": 617.8, "completions/mean_length": 203.0408905029297, "completions/mean_terminated_length": 203.0408905029297, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.0296451660261179, "epoch": 4.186715641261096, "frac_reward_zero_std": 0.12916666977107524, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 701182269.0, "reward": 8.030961322784425, "reward_std": 0.3274936407804489, "rewards/combined_reward/mean": 8.030961084365845, "rewards/combined_reward/std": 2.4156662106513975, "step": 6840 }, { "clip_ratio/high_max": 0.002374862630676944, "clip_ratio/high_mean": 0.0014434771015658042, "clip_ratio/low_mean": 0.0010484812562935985, "clip_ratio/low_min": 0.00032334302328915026, "clip_ratio/region_mean": 0.002491958363680169, "completions/clipped_ratio": 0.0, "completions/max_length": 625.1, "completions/max_terminated_length": 625.1, "completions/mean_length": 195.30573577880858, "completions/mean_terminated_length": 195.30573577880858, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.029184779711067678, "epoch": 4.205081114172024, "frac_reward_zero_std": 0.16666667275130748, "grad_norm": 0.0054931640625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 704232403.0, "reward": 8.24869999885559, "reward_std": 0.30624160915613174, "rewards/combined_reward/mean": 8.24869990348816, "rewards/combined_reward/std": 2.171760880947113, "step": 6870 }, { "clip_ratio/high_max": 0.0021462689401232637, "clip_ratio/high_mean": 0.0011441739894507918, "clip_ratio/low_mean": 0.000869812869617211, "clip_ratio/low_min": 0.00024717109069266976, "clip_ratio/region_mean": 0.0020139868844125885, "completions/clipped_ratio": 0.0, "completions/max_length": 694.4, "completions/max_terminated_length": 694.4, "completions/mean_length": 197.81432647705077, "completions/mean_terminated_length": 197.81432647705077, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.026130128232762218, "epoch": 4.22344658708295, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 707292170.0, "reward": 8.014334535598755, "reward_std": 0.3534664124250412, "rewards/combined_reward/mean": 8.014334344863892, "rewards/combined_reward/std": 2.3026664018630982, "step": 6900 }, { "clip_ratio/high_max": 0.0023502709863047738, "clip_ratio/high_mean": 0.0013325436040759087, "clip_ratio/low_mean": 0.0008391895737683323, "clip_ratio/low_min": 0.00022421649749351975, "clip_ratio/region_mean": 0.002171733186696656, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 196.2710968017578, "completions/mean_terminated_length": 196.2710968017578, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.028605909459292887, "epoch": 4.241812059993878, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 710346011.0, "reward": 8.000658702850341, "reward_std": 0.32322747707366944, "rewards/combined_reward/mean": 8.00065836906433, "rewards/combined_reward/std": 2.1348575830459593, "step": 6930 }, { "clip_ratio/high_max": 0.00267291635464062, "clip_ratio/high_mean": 0.0014667818885451803, "clip_ratio/low_mean": 0.0009962158408598043, "clip_ratio/low_min": 0.0002770037455775309, "clip_ratio/region_mean": 0.0024629977488075383, "completions/clipped_ratio": 0.0, "completions/max_length": 603.7, "completions/max_terminated_length": 603.7, "completions/mean_length": 202.51224517822266, "completions/mean_terminated_length": 202.51224517822266, "completions/min_length": 45.3, "completions/min_terminated_length": 45.3, "entropy": 0.033068387070670724, "epoch": 4.260177532904805, "frac_reward_zero_std": 0.13333333730697633, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 713423818.0, "reward": 8.07092514038086, "reward_std": 0.3550621777772903, "rewards/combined_reward/mean": 8.07092490196228, "rewards/combined_reward/std": 2.252958858013153, "step": 6960 }, { "clip_ratio/high_max": 0.0023092554770604086, "clip_ratio/high_mean": 0.0013080772620014613, "clip_ratio/low_mean": 0.0009737672581347094, "clip_ratio/low_min": 0.00030082996527198703, "clip_ratio/region_mean": 0.002281844494670319, "completions/clipped_ratio": 0.0, "completions/max_length": 547.3, "completions/max_terminated_length": 547.3, "completions/mean_length": 197.3882858276367, "completions/mean_terminated_length": 197.3882858276367, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.030096219495559733, "epoch": 4.278543005815733, "frac_reward_zero_std": 0.2000000037252903, "grad_norm": 0.00634765625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 716481949.0, "reward": 8.182129001617431, "reward_std": 0.3046834960579872, "rewards/combined_reward/mean": 8.182128667831421, "rewards/combined_reward/std": 2.0078811407089234, "step": 6990 }, { "clip_ratio/high_max": 0.0019904079075786286, "clip_ratio/high_mean": 0.0011768465220181193, "clip_ratio/low_mean": 0.001051372764959524, "clip_ratio/low_min": 0.0003344174462351172, "clip_ratio/region_mean": 0.00222821928812967, "completions/clipped_ratio": 0.0, "completions/max_length": 584.2, "completions/max_terminated_length": 584.2, "completions/mean_length": 196.54584197998048, "completions/mean_terminated_length": 196.54584197998048, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.027815381235753497, "epoch": 4.29690847872666, "frac_reward_zero_std": 0.19583333991467952, "grad_norm": 0.00677490234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 719536845.0, "reward": 8.052365493774413, "reward_std": 0.2901612862944603, "rewards/combined_reward/mean": 8.052365493774413, "rewards/combined_reward/std": 2.3157843708992005, "step": 7020 }, { "clip_ratio/high_max": 0.0033818735338475864, "clip_ratio/high_mean": 0.0019587125229008962, "clip_ratio/low_mean": 0.001155179446989981, "clip_ratio/low_min": 0.0003139207988472966, "clip_ratio/region_mean": 0.003113891975469111, "completions/clipped_ratio": 0.0, "completions/max_length": 622.5, "completions/max_terminated_length": 622.5, "completions/mean_length": 196.76120300292968, "completions/mean_terminated_length": 196.76120300292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.04162467539620896, "epoch": 4.315273951637588, "frac_reward_zero_std": 0.2250000059604645, "grad_norm": 0.005706787109375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 722592568.0, "reward": 8.028517389297486, "reward_std": 0.2955907240509987, "rewards/combined_reward/mean": 8.028516912460328, "rewards/combined_reward/std": 2.338058316707611, "step": 7050 }, { "clip_ratio/high_max": 0.002250416046687557, "clip_ratio/high_mean": 0.0013713432046642993, "clip_ratio/low_mean": 0.0009134167263255222, "clip_ratio/low_min": 0.00018598261716154714, "clip_ratio/region_mean": 0.002284759935961726, "completions/clipped_ratio": 0.0, "completions/max_length": 705.6, "completions/max_terminated_length": 705.6, "completions/mean_length": 201.8221420288086, "completions/mean_terminated_length": 201.8221420288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02923236615024507, "epoch": 4.333639424548515, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 725667725.0, "reward": 8.06669840812683, "reward_std": 0.34546761959791183, "rewards/combined_reward/mean": 8.066698169708252, "rewards/combined_reward/std": 2.2644297599792482, "step": 7080 }, { "clip_ratio/high_max": 0.0027395635891783363, "clip_ratio/high_mean": 0.0014890519426747535, "clip_ratio/low_mean": 0.0011196677437207351, "clip_ratio/low_min": 0.00027120542072225364, "clip_ratio/region_mean": 0.0026087196868805527, "completions/clipped_ratio": 0.0, "completions/max_length": 614.7, "completions/max_terminated_length": 614.7, "completions/mean_length": 199.66380767822267, "completions/mean_terminated_length": 199.66380767822267, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.03160299933515489, "epoch": 4.352004897459443, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 728734594.0, "reward": 8.202951622009277, "reward_std": 0.297206449508667, "rewards/combined_reward/mean": 8.202951240539551, "rewards/combined_reward/std": 2.1351240992546083, "step": 7110 }, { "clip_ratio/high_max": 0.0020226388861677453, "clip_ratio/high_mean": 0.001171319651560528, "clip_ratio/low_mean": 0.0009152591988822678, "clip_ratio/low_min": 0.00023183850474500407, "clip_ratio/region_mean": 0.002086578853292546, "completions/clipped_ratio": 0.0, "completions/max_length": 704.5, "completions/max_terminated_length": 704.5, "completions/mean_length": 202.40781860351564, "completions/mean_terminated_length": 202.40781860351564, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.027015909700033567, "epoch": 4.37037037037037, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 731812000.0, "reward": 8.246410751342774, "reward_std": 0.35009112656116487, "rewards/combined_reward/mean": 8.246410417556763, "rewards/combined_reward/std": 2.0817598581314085, "step": 7140 }, { "clip_ratio/high_max": 0.002575656987998324, "clip_ratio/high_mean": 0.0014859470099812218, "clip_ratio/low_mean": 0.000848440415696435, "clip_ratio/low_min": 0.0001661539824757104, "clip_ratio/region_mean": 0.0023343874335599443, "completions/clipped_ratio": 0.0, "completions/max_length": 508.1, "completions/max_terminated_length": 508.1, "completions/mean_length": 192.00130767822264, "completions/mean_terminated_length": 192.00130767822264, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02833788323526581, "epoch": 4.388735843281298, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 734849445.0, "reward": 8.26920280456543, "reward_std": 0.26557355746626854, "rewards/combined_reward/mean": 8.269202327728271, "rewards/combined_reward/std": 2.0512235283851625, "step": 7170 }, { "clip_ratio/high_max": 0.002289508156051549, "clip_ratio/high_mean": 0.001442379135551164, "clip_ratio/low_mean": 0.0008013310252863448, "clip_ratio/low_min": 0.00015544040191646977, "clip_ratio/region_mean": 0.0022437101802400623, "completions/clipped_ratio": 0.0, "completions/max_length": 606.5, "completions/max_terminated_length": 606.5, "completions/mean_length": 189.11042022705078, "completions/mean_terminated_length": 189.11042022705078, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.02517825740699967, "epoch": 4.407101316192225, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 737875789.0, "reward": 8.309944820404052, "reward_std": 0.36528085470199584, "rewards/combined_reward/mean": 8.309944295883179, "rewards/combined_reward/std": 1.8488682508468628, "step": 7200 }, { "clip_ratio/high_max": 0.00248465914822494, "clip_ratio/high_mean": 0.0014156900625190852, "clip_ratio/low_mean": 0.0009684123751261116, "clip_ratio/low_min": 0.0002823837712639943, "clip_ratio/region_mean": 0.002384102425033537, "completions/clipped_ratio": 0.0, "completions/max_length": 638.2, "completions/max_terminated_length": 638.2, "completions/mean_length": 199.6018280029297, "completions/mean_terminated_length": 199.6018280029297, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.029495792766101658, "epoch": 4.425466789103153, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 740942420.0, "reward": 7.86834397315979, "reward_std": 0.3272093623876572, "rewards/combined_reward/mean": 7.868343782424927, "rewards/combined_reward/std": 2.5620327472686766, "step": 7230 }, { "clip_ratio/high_max": 0.0023999042949678067, "clip_ratio/high_mean": 0.0013360954386977635, "clip_ratio/low_mean": 0.0009135353543873255, "clip_ratio/low_min": 0.00021990963068674318, "clip_ratio/region_mean": 0.0022496307981782595, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 198.4221420288086, "completions/mean_terminated_length": 198.4221420288086, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02975050937384367, "epoch": 4.44383226201408, "frac_reward_zero_std": 0.17500000707805158, "grad_norm": 0.0106201171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 744004521.0, "reward": 7.995103168487549, "reward_std": 0.3346598595380783, "rewards/combined_reward/mean": 7.9951029300689695, "rewards/combined_reward/std": 2.406872582435608, "step": 7260 }, { "clip_ratio/high_max": 0.00228469195208163, "clip_ratio/high_mean": 0.0014166627880816427, "clip_ratio/low_mean": 0.001070396035174781, "clip_ratio/low_min": 0.00037699440824023135, "clip_ratio/region_mean": 0.002487058821740599, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 197.0963592529297, "completions/mean_terminated_length": 197.0963592529297, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029692838573828338, "epoch": 4.462197734925008, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.007293701171875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 747061531.0, "reward": 7.881950187683105, "reward_std": 0.3445219546556473, "rewards/combined_reward/mean": 7.881950330734253, "rewards/combined_reward/std": 2.450170600414276, "step": 7290 }, { "clip_ratio/high_max": 0.0027769045994015568, "clip_ratio/high_mean": 0.0014369356003347397, "clip_ratio/low_mean": 0.0009550851297414434, "clip_ratio/low_min": 0.00022149562646518462, "clip_ratio/region_mean": 0.0023920207444461995, "completions/clipped_ratio": 0.0, "completions/max_length": 589.5, "completions/max_terminated_length": 589.5, "completions/mean_length": 200.45390930175782, "completions/mean_terminated_length": 200.45390930175782, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.030569868441671134, "epoch": 4.480563207835935, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 750131434.0, "reward": 8.204556798934936, "reward_std": 0.3541791394352913, "rewards/combined_reward/mean": 8.204556512832642, "rewards/combined_reward/std": 2.2757139563560487, "step": 7320 }, { "clip_ratio/high_max": 0.002426535653648898, "clip_ratio/high_mean": 0.001380598621350752, "clip_ratio/low_mean": 0.0008399231952353148, "clip_ratio/low_min": 0.00017722954992981007, "clip_ratio/region_mean": 0.002220521796940981, "completions/clipped_ratio": 0.0, "completions/max_length": 599.3, "completions/max_terminated_length": 599.3, "completions/mean_length": 185.46693267822266, "completions/mean_terminated_length": 185.46693267822266, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.028509683425848684, "epoch": 4.498928680746863, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.0257568359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 753143787.0, "reward": 7.982299327850342, "reward_std": 0.35472443103790285, "rewards/combined_reward/mean": 7.982299137115478, "rewards/combined_reward/std": 2.375264608860016, "step": 7350 }, { "clip_ratio/high_max": 0.0027815997433208394, "clip_ratio/high_mean": 0.0016473406263685319, "clip_ratio/low_mean": 0.0011931670024447763, "clip_ratio/low_min": 0.00035369252606566685, "clip_ratio/region_mean": 0.0028405076547642237, "completions/clipped_ratio": 0.0, "completions/max_length": 614.3, "completions/max_terminated_length": 614.3, "completions/mean_length": 202.99036865234376, "completions/mean_terminated_length": 202.99036865234376, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.03540386447372536, "epoch": 4.51729415365779, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 756223430.0, "reward": 7.884941864013672, "reward_std": 0.2962191145867109, "rewards/combined_reward/mean": 7.884941434860229, "rewards/combined_reward/std": 2.5361255288124083, "step": 7380 }, { "clip_ratio/high_max": 0.002142202110553626, "clip_ratio/high_mean": 0.0012426315410266398, "clip_ratio/low_mean": 0.0008558733574318466, "clip_ratio/low_min": 0.00019594716941355727, "clip_ratio/region_mean": 0.002098504911797742, "completions/clipped_ratio": 0.0, "completions/max_length": 563.7, "completions/max_terminated_length": 563.7, "completions/mean_length": 198.96562957763672, "completions/mean_terminated_length": 198.96562957763672, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.03008669289605071, "epoch": 4.535659626568718, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 759287618.0, "reward": 8.30747127532959, "reward_std": 0.3003626450896263, "rewards/combined_reward/mean": 8.307470703125, "rewards/combined_reward/std": 1.9623859763145446, "step": 7410 }, { "clip_ratio/high_max": 0.00250838162949852, "clip_ratio/high_mean": 0.0014328367561878017, "clip_ratio/low_mean": 0.001029680506326258, "clip_ratio/low_min": 0.0003302029998546156, "clip_ratio/region_mean": 0.0024625172702750814, "completions/clipped_ratio": 0.0, "completions/max_length": 701.5, "completions/max_terminated_length": 701.5, "completions/mean_length": 206.22786865234374, "completions/mean_terminated_length": 206.22786865234374, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028949435614049435, "epoch": 4.554025099479645, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 762379693.0, "reward": 8.029173040390015, "reward_std": 0.40383909940719603, "rewards/combined_reward/mean": 8.029172563552857, "rewards/combined_reward/std": 2.2662413120269775, "step": 7440 }, { "clip_ratio/high_max": 0.0023252251203909205, "clip_ratio/high_mean": 0.0013675946579799832, "clip_ratio/low_mean": 0.0010023198932079442, "clip_ratio/low_min": 0.00029401312300857775, "clip_ratio/region_mean": 0.0023699145516729915, "completions/clipped_ratio": 0.0, "completions/max_length": 644.4, "completions/max_terminated_length": 644.4, "completions/mean_length": 201.24297332763672, "completions/mean_terminated_length": 201.24297332763672, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.0296476477291435, "epoch": 4.572390572390573, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.00421142578125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 765452626.0, "reward": 8.078011989593506, "reward_std": 0.3378586396574974, "rewards/combined_reward/mean": 8.078011465072631, "rewards/combined_reward/std": 2.0488444685935976, "step": 7470 }, { "clip_ratio/high_max": 0.002422533626668155, "clip_ratio/high_mean": 0.0013119089576017966, "clip_ratio/low_mean": 0.0008482709896876864, "clip_ratio/low_min": 0.00022875838258187286, "clip_ratio/region_mean": 0.0021601799366180786, "completions/clipped_ratio": 0.0, "completions/max_length": 626.5, "completions/max_terminated_length": 626.5, "completions/mean_length": 201.5033920288086, "completions/mean_terminated_length": 201.5033920288086, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.025770229916088284, "epoch": 4.5907560453015, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 768526559.0, "reward": 8.17113642692566, "reward_std": 0.36850923001766206, "rewards/combined_reward/mean": 8.171136236190796, "rewards/combined_reward/std": 2.1874457955360413, "step": 7500 }, { "clip_ratio/high_max": 0.0023193219138192946, "clip_ratio/high_mean": 0.0014386832233867608, "clip_ratio/low_mean": 0.0009302306653504881, "clip_ratio/low_min": 0.00020585102514208604, "clip_ratio/region_mean": 0.002368913854297716, "completions/clipped_ratio": 0.0, "completions/max_length": 623.2, "completions/max_terminated_length": 623.2, "completions/mean_length": 204.23411865234374, "completions/mean_terminated_length": 204.23411865234374, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02757537238455067, "epoch": 4.609121518212428, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 771610978.0, "reward": 8.341620349884034, "reward_std": 0.3448057159781456, "rewards/combined_reward/mean": 8.341620206832886, "rewards/combined_reward/std": 1.922659945487976, "step": 7530 }, { "clip_ratio/high_max": 0.0022953462961595506, "clip_ratio/high_mean": 0.0013765656909527023, "clip_ratio/low_mean": 0.0011713262952980585, "clip_ratio/low_min": 0.00044961450451713367, "clip_ratio/region_mean": 0.0025478920181437087, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 199.8674514770508, "completions/mean_terminated_length": 199.8674514770508, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.03205722781519095, "epoch": 4.627486991123355, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 774678629.0, "reward": 7.8421741962432865, "reward_std": 0.3213977813720703, "rewards/combined_reward/mean": 7.842173719406128, "rewards/combined_reward/std": 2.400368881225586, "step": 7560 }, { "clip_ratio/high_max": 0.0025990853100665845, "clip_ratio/high_mean": 0.0014762717482274942, "clip_ratio/low_mean": 0.0009001361244979004, "clip_ratio/low_min": 0.00023438301723217593, "clip_ratio/region_mean": 0.002376407893219342, "completions/clipped_ratio": 0.0, "completions/max_length": 548.9, "completions/max_terminated_length": 548.9, "completions/mean_length": 198.61042327880858, "completions/mean_terminated_length": 198.61042327880858, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03182392034990092, "epoch": 4.645852464034283, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 777741453.0, "reward": 7.886397171020508, "reward_std": 0.30338632836937907, "rewards/combined_reward/mean": 7.886396741867065, "rewards/combined_reward/std": 2.3295358180999757, "step": 7590 }, { "clip_ratio/high_max": 0.0024246236163889988, "clip_ratio/high_mean": 0.0014298507961939323, "clip_ratio/low_mean": 0.000990616372776761, "clip_ratio/low_min": 0.0002746435976102172, "clip_ratio/region_mean": 0.002420467175155257, "completions/clipped_ratio": 0.0, "completions/max_length": 508.9, "completions/max_terminated_length": 508.9, "completions/mean_length": 189.61276550292968, "completions/mean_terminated_length": 189.61276550292968, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.02821761022011439, "epoch": 4.66421793694521, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.003509521484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 780769726.0, "reward": 8.130964708328246, "reward_std": 0.3186579279601574, "rewards/combined_reward/mean": 8.130964422225953, "rewards/combined_reward/std": 2.1957751512527466, "step": 7620 }, { "clip_ratio/high_max": 0.002583761579201867, "clip_ratio/high_mean": 0.0014734674177210158, "clip_ratio/low_mean": 0.0010389625952181329, "clip_ratio/low_min": 0.00028731104830512775, "clip_ratio/region_mean": 0.00251243000869484, "completions/clipped_ratio": 0.0, "completions/max_length": 695.5, "completions/max_terminated_length": 695.5, "completions/mean_length": 217.6737045288086, "completions/mean_terminated_length": 217.6737045288086, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.030876279482617973, "epoch": 4.682583409856137, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 783905753.0, "reward": 8.000315618515014, "reward_std": 0.3499706745147705, "rewards/combined_reward/mean": 8.000315284729004, "rewards/combined_reward/std": 2.4130021214485167, "step": 7650 }, { "clip_ratio/high_max": 0.002427976648323238, "clip_ratio/high_mean": 0.001373708271664024, "clip_ratio/low_mean": 0.0008955881015936029, "clip_ratio/low_min": 0.0002744933677604422, "clip_ratio/region_mean": 0.002269296391265622, "completions/clipped_ratio": 0.0, "completions/max_length": 723.3, "completions/max_terminated_length": 723.3, "completions/mean_length": 201.37162017822266, "completions/mean_terminated_length": 201.37162017822266, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.029555653408169746, "epoch": 4.700948882767064, "frac_reward_zero_std": 0.15833333916962147, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 786979180.0, "reward": 8.047964763641357, "reward_std": 0.35227497965097426, "rewards/combined_reward/mean": 8.047964811325073, "rewards/combined_reward/std": 2.200173056125641, "step": 7680 }, { "clip_ratio/high_max": 0.0028745211981004103, "clip_ratio/high_mean": 0.0017134810511682493, "clip_ratio/low_mean": 0.0009975179497511515, "clip_ratio/low_min": 0.0003240416898430946, "clip_ratio/region_mean": 0.002710998994007241, "completions/clipped_ratio": 0.0, "completions/max_length": 710.9, "completions/max_terminated_length": 710.9, "completions/mean_length": 209.1007843017578, "completions/mean_terminated_length": 209.1007843017578, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.031448728994776805, "epoch": 4.719314355677992, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 790082287.0, "reward": 7.817307472229004, "reward_std": 0.35689702332019807, "rewards/combined_reward/mean": 7.817307424545288, "rewards/combined_reward/std": 2.6758187413215637, "step": 7710 }, { "clip_ratio/high_max": 0.0029003335497691297, "clip_ratio/high_mean": 0.0017664469664547748, "clip_ratio/low_mean": 0.0010104160815293047, "clip_ratio/low_min": 0.00023104109471508612, "clip_ratio/region_mean": 0.002776863050530665, "completions/clipped_ratio": 0.0, "completions/max_length": 545.8, "completions/max_terminated_length": 545.8, "completions/mean_length": 205.55130920410156, "completions/mean_terminated_length": 205.55130920410156, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.031626573158428074, "epoch": 4.737679828588919, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 793171764.0, "reward": 8.008633422851563, "reward_std": 0.4170594930648804, "rewards/combined_reward/mean": 8.008633232116699, "rewards/combined_reward/std": 2.4235588669776917, "step": 7740 }, { "clip_ratio/high_max": 0.002769049420506538, "clip_ratio/high_mean": 0.0016481446532755702, "clip_ratio/low_mean": 0.0008664570856732704, "clip_ratio/low_min": 0.000285301621867499, "clip_ratio/region_mean": 0.002514601737736181, "completions/clipped_ratio": 0.0, "completions/max_length": 720.2, "completions/max_terminated_length": 720.2, "completions/mean_length": 215.89922332763672, "completions/mean_terminated_length": 215.89922332763672, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.0330362736635531, "epoch": 4.756045301499847, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 796300977.0, "reward": 7.82343692779541, "reward_std": 0.37996661365032197, "rewards/combined_reward/mean": 7.82343635559082, "rewards/combined_reward/std": 2.474334442615509, "step": 7770 }, { "clip_ratio/high_max": 0.002660003643541131, "clip_ratio/high_mean": 0.001574029140950491, "clip_ratio/low_mean": 0.0012250491439772305, "clip_ratio/low_min": 0.000400967419166894, "clip_ratio/region_mean": 0.0027990782778942957, "completions/clipped_ratio": 0.0, "completions/max_length": 613.7, "completions/max_terminated_length": 613.7, "completions/mean_length": 203.67917175292968, "completions/mean_terminated_length": 203.67917175292968, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.03579608077804248, "epoch": 4.774410774410774, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 799383265.0, "reward": 7.783587884902954, "reward_std": 0.3767504900693893, "rewards/combined_reward/mean": 7.783587503433227, "rewards/combined_reward/std": 2.4410789012908936, "step": 7800 }, { "clip_ratio/high_max": 0.002758466449934834, "clip_ratio/high_mean": 0.0016069311551594487, "clip_ratio/low_mean": 0.0010110308906102242, "clip_ratio/low_min": 0.00021937763206854774, "clip_ratio/region_mean": 0.002617962038978779, "completions/clipped_ratio": 0.0, "completions/max_length": 640.9, "completions/max_terminated_length": 640.9, "completions/mean_length": 196.8898483276367, "completions/mean_terminated_length": 196.8898483276367, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.03325985143892467, "epoch": 4.792776247321702, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 802439482.0, "reward": 8.063335371017455, "reward_std": 0.36622300893068316, "rewards/combined_reward/mean": 8.063335084915161, "rewards/combined_reward/std": 2.097376596927643, "step": 7830 }, { "clip_ratio/high_max": 0.0024465918652519275, "clip_ratio/high_mean": 0.0014672546691144816, "clip_ratio/low_mean": 0.0011271271818259266, "clip_ratio/low_min": 0.00041174960182009576, "clip_ratio/region_mean": 0.0025943818657348555, "completions/clipped_ratio": 0.0, "completions/max_length": 581.3, "completions/max_terminated_length": 581.3, "completions/mean_length": 202.20938110351562, "completions/mean_terminated_length": 202.20938110351562, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.03345395714665453, "epoch": 4.811141720232629, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 805516126.0, "reward": 8.01763162612915, "reward_std": 0.3647181734442711, "rewards/combined_reward/mean": 8.017631530761719, "rewards/combined_reward/std": 2.302188491821289, "step": 7860 }, { "clip_ratio/high_max": 0.0024940535532853875, "clip_ratio/high_mean": 0.0014916746401771282, "clip_ratio/low_mean": 0.0010362092920937964, "clip_ratio/low_min": 0.000331896807620069, "clip_ratio/region_mean": 0.002527883921963318, "completions/clipped_ratio": 0.0, "completions/max_length": 546.3, "completions/max_terminated_length": 546.3, "completions/mean_length": 199.88412170410157, "completions/mean_terminated_length": 199.88412170410157, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.030730843730270862, "epoch": 4.829507193143557, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 808583841.0, "reward": 7.715697383880615, "reward_std": 0.41294152587652205, "rewards/combined_reward/mean": 7.715697193145752, "rewards/combined_reward/std": 2.5318157076835632, "step": 7890 }, { "clip_ratio/high_max": 0.002683554670754044, "clip_ratio/high_mean": 0.0015291253462540529, "clip_ratio/low_mean": 0.0009807013174092086, "clip_ratio/low_min": 0.00026406375861067014, "clip_ratio/region_mean": 0.002509826683672145, "completions/clipped_ratio": 0.0, "completions/max_length": 555.1, "completions/max_terminated_length": 555.1, "completions/mean_length": 204.9440170288086, "completions/mean_terminated_length": 204.9440170288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.031133974095185597, "epoch": 4.847872666054484, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 811670986.0, "reward": 7.975838279724121, "reward_std": 0.3762900710105896, "rewards/combined_reward/mean": 7.97583794593811, "rewards/combined_reward/std": 2.184993231296539, "step": 7920 }, { "clip_ratio/high_max": 0.002750430195980395, "clip_ratio/high_mean": 0.0015824996723191969, "clip_ratio/low_mean": 0.00110809738386403, "clip_ratio/low_min": 0.00035830589088921744, "clip_ratio/region_mean": 0.0026905970473308114, "completions/clipped_ratio": 0.0, "completions/max_length": 592.8, "completions/max_terminated_length": 592.8, "completions/mean_length": 196.5510467529297, "completions/mean_terminated_length": 196.5510467529297, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03351377487803499, "epoch": 4.866238138965412, "frac_reward_zero_std": 0.17500000074505806, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 814725902.0, "reward": 7.990303039550781, "reward_std": 0.3996566280722618, "rewards/combined_reward/mean": 7.990302562713623, "rewards/combined_reward/std": 2.3205293893814085, "step": 7950 }, { "clip_ratio/high_max": 0.002218148768830967, "clip_ratio/high_mean": 0.0012360845539660659, "clip_ratio/low_mean": 0.001086082302693588, "clip_ratio/low_min": 0.0003141527915431652, "clip_ratio/region_mean": 0.002322166861752824, "completions/clipped_ratio": 0.0, "completions/max_length": 579.1, "completions/max_terminated_length": 579.1, "completions/mean_length": 209.36693267822267, "completions/mean_terminated_length": 209.36693267822267, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.027950360424195728, "epoch": 4.884603611876339, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 817830031.0, "reward": 7.907508611679077, "reward_std": 0.3433987110853195, "rewards/combined_reward/mean": 7.9075085639953615, "rewards/combined_reward/std": 2.3119752764701844, "step": 7980 }, { "clip_ratio/high_max": 0.0023818729134897393, "clip_ratio/high_mean": 0.0013797868328159287, "clip_ratio/low_mean": 0.0009711140963797031, "clip_ratio/low_min": 0.0002545367909381942, "clip_ratio/region_mean": 0.0023509009270734774, "completions/clipped_ratio": 0.0, "completions/max_length": 598.5, "completions/max_terminated_length": 598.5, "completions/mean_length": 202.72995300292968, "completions/mean_terminated_length": 202.72995300292968, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.029833206627517938, "epoch": 4.902969084787267, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.00933837890625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 820908674.0, "reward": 7.656227064132691, "reward_std": 0.3700548678636551, "rewards/combined_reward/mean": 7.656226873397827, "rewards/combined_reward/std": 2.4980942249298095, "step": 8010 }, { "clip_ratio/high_max": 0.0023775969428243117, "clip_ratio/high_mean": 0.0013622956772451288, "clip_ratio/low_mean": 0.0009215283072990133, "clip_ratio/low_min": 0.00027470774972850146, "clip_ratio/region_mean": 0.002283823975691727, "completions/clipped_ratio": 0.0, "completions/max_length": 611.6, "completions/max_terminated_length": 611.6, "completions/mean_length": 196.6869873046875, "completions/mean_terminated_length": 196.6869873046875, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.029964654023448625, "epoch": 4.921334557698194, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 823964112.0, "reward": 8.14067063331604, "reward_std": 0.3750671029090881, "rewards/combined_reward/mean": 8.140670204162598, "rewards/combined_reward/std": 2.2861721754074096, "step": 8040 }, { "clip_ratio/high_max": 0.002941379991049568, "clip_ratio/high_mean": 0.0016491410079955435, "clip_ratio/low_mean": 0.0010616152191990599, "clip_ratio/low_min": 0.00033866760713863187, "clip_ratio/region_mean": 0.002710756228285997, "completions/clipped_ratio": 0.0, "completions/max_length": 673.6, "completions/max_terminated_length": 673.6, "completions/mean_length": 200.74323577880858, "completions/mean_terminated_length": 200.74323577880858, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03379698609933257, "epoch": 4.939700030609122, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0150146484375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 827035126.0, "reward": 7.724186420440674, "reward_std": 0.384482279419899, "rewards/combined_reward/mean": 7.724185991287231, "rewards/combined_reward/std": 2.5504367113113404, "step": 8070 }, { "clip_ratio/high_max": 0.002616350239259191, "clip_ratio/high_mean": 0.001542312489861312, "clip_ratio/low_mean": 0.001051751531122136, "clip_ratio/low_min": 0.0002914807645235366, "clip_ratio/region_mean": 0.0025940640320186504, "completions/clipped_ratio": 0.0, "completions/max_length": 580.8, "completions/max_terminated_length": 580.8, "completions/mean_length": 201.10130615234374, "completions/mean_terminated_length": 201.10130615234374, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.03401536429300904, "epoch": 4.958065503520049, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 830107515.0, "reward": 7.875389146804809, "reward_std": 0.3661284863948822, "rewards/combined_reward/mean": 7.875388765335083, "rewards/combined_reward/std": 2.4502355098724364, "step": 8100 }, { "clip_ratio/high_max": 0.002568520763209866, "clip_ratio/high_mean": 0.0013747970129770692, "clip_ratio/low_mean": 0.001000224321736217, "clip_ratio/low_min": 0.00023753048371872866, "clip_ratio/region_mean": 0.0023750213338644246, "completions/clipped_ratio": 0.0, "completions/max_length": 538.8, "completions/max_terminated_length": 538.8, "completions/mean_length": 193.25729675292968, "completions/mean_terminated_length": 193.25729675292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028017692423115175, "epoch": 4.976430976430977, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 833149783.0, "reward": 7.958741283416748, "reward_std": 0.3571596175432205, "rewards/combined_reward/mean": 7.958741140365601, "rewards/combined_reward/std": 2.2766525864601137, "step": 8130 }, { "clip_ratio/high_max": 0.0029463576235381574, "clip_ratio/high_mean": 0.0016393387066879464, "clip_ratio/low_mean": 0.001194673346617492, "clip_ratio/low_min": 0.00027305127320384297, "clip_ratio/region_mean": 0.0028340120382684595, "completions/clipped_ratio": 0.0, "completions/max_length": 637.5, "completions/max_terminated_length": 637.5, "completions/mean_length": 201.22109985351562, "completions/mean_terminated_length": 201.22109985351562, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.033233796386048196, "epoch": 4.994796449341904, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.0098876953125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 836222632.0, "reward": 8.106711769104004, "reward_std": 0.3392436563968658, "rewards/combined_reward/mean": 8.10671133995056, "rewards/combined_reward/std": 2.134864163398743, "step": 8160 }, { "clip_ratio/high_max": 0.0025880082625840493, "clip_ratio/high_mean": 0.00148931759906838, "clip_ratio/low_mean": 0.0009075543715942146, "clip_ratio/low_min": 0.00032156467106256444, "clip_ratio/region_mean": 0.0023968719908872903, "completions/clipped_ratio": 0.0, "completions/max_length": 623.1, "completions/max_terminated_length": 623.1, "completions/mean_length": 202.61510772705077, "completions/mean_terminated_length": 202.61510772705077, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.03040630010477567, "epoch": 5.012855831037649, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 839300834.0, "reward": 8.068760538101197, "reward_std": 0.34978931248188017, "rewards/combined_reward/mean": 8.06875991821289, "rewards/combined_reward/std": 2.165126860141754, "step": 8190 }, { "clip_ratio/high_max": 0.0018653361020066464, "clip_ratio/high_mean": 0.001127519483634387, "clip_ratio/low_mean": 0.0007501003520398324, "clip_ratio/low_min": 0.0002158761310662764, "clip_ratio/region_mean": 0.0018776198300959853, "completions/clipped_ratio": 0.0, "completions/max_length": 636.4, "completions/max_terminated_length": 636.4, "completions/mean_length": 200.3440170288086, "completions/mean_terminated_length": 200.3440170288086, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.03013422912918031, "epoch": 5.031221303948577, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 842370315.0, "reward": 7.898865985870361, "reward_std": 0.34734165370464326, "rewards/combined_reward/mean": 7.8988659381866455, "rewards/combined_reward/std": 2.3973360657691956, "step": 8220 }, { "clip_ratio/high_max": 0.002240107329756332, "clip_ratio/high_mean": 0.0012809289827904044, "clip_ratio/low_mean": 0.0007695268035604385, "clip_ratio/low_min": 0.00020662514628687253, "clip_ratio/region_mean": 0.0020504557731328534, "completions/clipped_ratio": 0.0, "completions/max_length": 624.2, "completions/max_terminated_length": 624.2, "completions/mean_length": 187.50443115234376, "completions/mean_terminated_length": 187.50443115234376, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.0318083891334633, "epoch": 5.049586776859504, "frac_reward_zero_std": 0.1541666727513075, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 845390492.0, "reward": 7.8934910774230955, "reward_std": 0.31513611972332, "rewards/combined_reward/mean": 7.893490886688232, "rewards/combined_reward/std": 2.447055089473724, "step": 8250 }, { "clip_ratio/high_max": 0.0020592323892439407, "clip_ratio/high_mean": 0.0011292293190005391, "clip_ratio/low_mean": 0.0008795237157755764, "clip_ratio/low_min": 0.00023544347495771945, "clip_ratio/region_mean": 0.0020087530264087643, "completions/clipped_ratio": 0.0, "completions/max_length": 529.4, "completions/max_terminated_length": 529.4, "completions/mean_length": 185.32109832763672, "completions/mean_terminated_length": 185.32109832763672, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.029354873097812137, "epoch": 5.067952249770432, "frac_reward_zero_std": 0.2000000085681677, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 848402285.0, "reward": 8.175055694580077, "reward_std": 0.3234966218471527, "rewards/combined_reward/mean": 8.175055456161498, "rewards/combined_reward/std": 2.195690095424652, "step": 8280 }, { "clip_ratio/high_max": 0.0018791746648882206, "clip_ratio/high_mean": 0.001078641816881524, "clip_ratio/low_mean": 0.0007205240846815287, "clip_ratio/low_min": 0.00020751942744633804, "clip_ratio/region_mean": 0.0017991658967124143, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 198.75885772705078, "completions/mean_terminated_length": 198.75885772705078, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.026707711761506894, "epoch": 5.086317722681359, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 851465679.0, "reward": 8.209881734848022, "reward_std": 0.34610054194927214, "rewards/combined_reward/mean": 8.209881353378297, "rewards/combined_reward/std": 2.167846715450287, "step": 8310 }, { "clip_ratio/high_max": 0.002097723219776526, "clip_ratio/high_mean": 0.0012305676832814545, "clip_ratio/low_mean": 0.0007413383871607948, "clip_ratio/low_min": 0.0002416704286588356, "clip_ratio/region_mean": 0.001971906039883227, "completions/clipped_ratio": 0.0, "completions/max_length": 563.7, "completions/max_terminated_length": 563.7, "completions/mean_length": 201.96068115234374, "completions/mean_terminated_length": 201.96068115234374, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.027834513845543068, "epoch": 5.104683195592287, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 854541368.0, "reward": 8.253080320358276, "reward_std": 0.357078455388546, "rewards/combined_reward/mean": 8.253080463409423, "rewards/combined_reward/std": 2.0850369691848756, "step": 8340 }, { "clip_ratio/high_max": 0.0019288489798782392, "clip_ratio/high_mean": 0.0011452313540454876, "clip_ratio/low_mean": 0.0008640429873291093, "clip_ratio/low_min": 0.0002859963261774586, "clip_ratio/region_mean": 0.0020092743269439477, "completions/clipped_ratio": 0.0, "completions/max_length": 628.8, "completions/max_terminated_length": 628.8, "completions/mean_length": 202.6877670288086, "completions/mean_terminated_length": 202.6877670288086, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.02861202749578903, "epoch": 5.123048668503214, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 857619849.0, "reward": 8.138101720809937, "reward_std": 0.2875575624406338, "rewards/combined_reward/mean": 8.13810167312622, "rewards/combined_reward/std": 2.207552123069763, "step": 8370 }, { "clip_ratio/high_max": 0.002252952312119305, "clip_ratio/high_mean": 0.0012938848274643532, "clip_ratio/low_mean": 0.0008086586339535036, "clip_ratio/low_min": 0.0002550336425580705, "clip_ratio/region_mean": 0.0021025434708766017, "completions/clipped_ratio": 0.0, "completions/max_length": 652.7, "completions/max_terminated_length": 652.7, "completions/mean_length": 200.54297485351563, "completions/mean_terminated_length": 200.54297485351563, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.03365719044425835, "epoch": 5.141414141414141, "frac_reward_zero_std": 0.1500000014901161, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 860690094.0, "reward": 7.953445291519165, "reward_std": 0.3248763635754585, "rewards/combined_reward/mean": 7.953444862365723, "rewards/combined_reward/std": 2.3875218510627745, "step": 8400 }, { "clip_ratio/high_max": 0.002114932561138024, "clip_ratio/high_mean": 0.0012693094925149732, "clip_ratio/low_mean": 0.0008739599788289827, "clip_ratio/low_min": 0.00019764435540613098, "clip_ratio/region_mean": 0.002143269462006477, "completions/clipped_ratio": 0.0, "completions/max_length": 565.7, "completions/max_terminated_length": 565.7, "completions/mean_length": 205.1424530029297, "completions/mean_terminated_length": 205.1424530029297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.032804447179660204, "epoch": 5.159167431894705, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 863778001.0, "reward": 8.027925252914429, "reward_std": 0.4559808760881424, "rewards/combined_reward/mean": 8.02792501449585, "rewards/combined_reward/std": 2.315766620635986, "step": 8430 }, { "clip_ratio/high_max": 0.0020292421669485822, "clip_ratio/high_mean": 0.0011785069160396232, "clip_ratio/low_mean": 0.0008004027348457991, "clip_ratio/low_min": 0.0001931348653063954, "clip_ratio/region_mean": 0.00197890964676238, "completions/clipped_ratio": 0.0, "completions/max_length": 552.3, "completions/max_terminated_length": 552.3, "completions/mean_length": 191.77448577880858, "completions/mean_terminated_length": 191.77448577880858, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028112400183454157, "epoch": 5.177532904805632, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 866814575.0, "reward": 8.128009033203124, "reward_std": 0.3161967918276787, "rewards/combined_reward/mean": 8.12800874710083, "rewards/combined_reward/std": 2.262754571437836, "step": 8460 }, { "clip_ratio/high_max": 0.0019453766367708643, "clip_ratio/high_mean": 0.0011116384680159778, "clip_ratio/low_mean": 0.0006326500909684304, "clip_ratio/low_min": 0.00015473809786878216, "clip_ratio/region_mean": 0.001744288556559089, "completions/clipped_ratio": 0.0, "completions/max_length": 606.5, "completions/max_terminated_length": 606.5, "completions/mean_length": 187.5237014770508, "completions/mean_terminated_length": 187.5237014770508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028572945793469747, "epoch": 5.19589837771656, "frac_reward_zero_std": 0.2041666716337204, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 869834826.0, "reward": 8.045674562454224, "reward_std": 0.3688571184873581, "rewards/combined_reward/mean": 8.045674085617065, "rewards/combined_reward/std": 2.281425416469574, "step": 8490 }, { "clip_ratio/high_max": 0.0022466197437703765, "clip_ratio/high_mean": 0.0011566876504123986, "clip_ratio/low_mean": 0.0007979900887827777, "clip_ratio/low_min": 0.0002192854701813, "clip_ratio/region_mean": 0.001954677749502783, "completions/clipped_ratio": 0.0, "completions/max_length": 651.9, "completions/max_terminated_length": 651.9, "completions/mean_length": 195.0104202270508, "completions/mean_terminated_length": 195.0104202270508, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.031080921087414025, "epoch": 5.214263850627487, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0269775390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 872883826.0, "reward": 7.788765335083008, "reward_std": 0.3564063623547554, "rewards/combined_reward/mean": 7.788765144348145, "rewards/combined_reward/std": 2.4557645201683043, "step": 8520 }, { "clip_ratio/high_max": 0.0019032412325032055, "clip_ratio/high_mean": 0.0011189699711394497, "clip_ratio/low_mean": 0.0007590339169837535, "clip_ratio/low_min": 0.00029513856788980773, "clip_ratio/region_mean": 0.0018780039084958845, "completions/clipped_ratio": 0.0, "completions/max_length": 664.1, "completions/max_terminated_length": 664.1, "completions/mean_length": 199.73906860351562, "completions/mean_terminated_length": 199.73906860351562, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.02889129358033339, "epoch": 5.232629323538414, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.01275634765625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 875950984.0, "reward": 8.217450952529907, "reward_std": 0.30023142471909525, "rewards/combined_reward/mean": 8.217450857162476, "rewards/combined_reward/std": 2.202113461494446, "step": 8550 }, { "clip_ratio/high_max": 0.0020959269864154824, "clip_ratio/high_mean": 0.0011577912679058499, "clip_ratio/low_mean": 0.0007090042397370174, "clip_ratio/low_min": 0.00014736591062198084, "clip_ratio/region_mean": 0.0018667955033985588, "completions/clipped_ratio": 0.0, "completions/max_length": 588.3, "completions/max_terminated_length": 588.3, "completions/mean_length": 192.7971389770508, "completions/mean_terminated_length": 192.7971389770508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029914426694934566, "epoch": 5.2509947964493415, "frac_reward_zero_std": 0.18750000782310963, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 878991485.0, "reward": 7.77699728012085, "reward_std": 0.44613738507032397, "rewards/combined_reward/mean": 7.77699728012085, "rewards/combined_reward/std": 2.5041526436805723, "step": 8580 }, { "clip_ratio/high_max": 0.0021025202125504925, "clip_ratio/high_mean": 0.001144776405514373, "clip_ratio/low_mean": 0.0008431482982511322, "clip_ratio/low_min": 0.0002294234497336826, "clip_ratio/region_mean": 0.001987924698429803, "completions/clipped_ratio": 0.0, "completions/max_length": 524.9, "completions/max_terminated_length": 524.9, "completions/mean_length": 182.5773483276367, "completions/mean_terminated_length": 182.5773483276367, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.031176976502562563, "epoch": 5.269360269360269, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 881992742.0, "reward": 8.10648250579834, "reward_std": 0.3705909952521324, "rewards/combined_reward/mean": 8.106482315063477, "rewards/combined_reward/std": 2.328423869609833, "step": 8610 }, { "clip_ratio/high_max": 0.0022426934267665882, "clip_ratio/high_mean": 0.0013597029724526995, "clip_ratio/low_mean": 0.0008239572744059842, "clip_ratio/low_min": 0.00023612882869201713, "clip_ratio/region_mean": 0.0021836602599554075, "completions/clipped_ratio": 0.0, "completions/max_length": 578.3, "completions/max_terminated_length": 578.3, "completions/mean_length": 202.5299530029297, "completions/mean_terminated_length": 202.5299530029297, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03149629245356967, "epoch": 5.2877257422711965, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 885070617.0, "reward": 7.924777841567993, "reward_std": 0.4056656926870346, "rewards/combined_reward/mean": 7.924777603149414, "rewards/combined_reward/std": 2.2659695625305174, "step": 8640 }, { "clip_ratio/high_max": 0.002138426576857455, "clip_ratio/high_mean": 0.0012525133640641191, "clip_ratio/low_mean": 0.0008449912224023137, "clip_ratio/low_min": 0.00025523650717029035, "clip_ratio/region_mean": 0.0020975045937423904, "completions/clipped_ratio": 0.0, "completions/max_length": 587.7, "completions/max_terminated_length": 587.7, "completions/mean_length": 195.2750030517578, "completions/mean_terminated_length": 195.2750030517578, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.03398564293359717, "epoch": 5.306091215182124, "frac_reward_zero_std": 0.12500000186264515, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 888120633.0, "reward": 8.0041184425354, "reward_std": 0.40263726115226744, "rewards/combined_reward/mean": 8.004117965698242, "rewards/combined_reward/std": 2.4381486654281614, "step": 8670 }, { "clip_ratio/high_max": 0.0019663314252587345, "clip_ratio/high_mean": 0.0010633114448864943, "clip_ratio/low_mean": 0.000713604311022209, "clip_ratio/low_min": 0.00024735414820800845, "clip_ratio/region_mean": 0.0017769157653674482, "completions/clipped_ratio": 0.0, "completions/max_length": 604.9, "completions/max_terminated_length": 604.9, "completions/mean_length": 190.52396392822266, "completions/mean_terminated_length": 190.52396392822266, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029402432699377336, "epoch": 5.3244566880930515, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 891152405.0, "reward": 7.7922508239746096, "reward_std": 0.3424156755208969, "rewards/combined_reward/mean": 7.7922505855560305, "rewards/combined_reward/std": 2.4047576189041138, "step": 8700 }, { "clip_ratio/high_max": 0.0020103737406316213, "clip_ratio/high_mean": 0.0011289595526856526, "clip_ratio/low_mean": 0.0007260395228513517, "clip_ratio/low_min": 0.00022373098108801058, "clip_ratio/region_mean": 0.0018549990796600468, "completions/clipped_ratio": 0.0, "completions/max_length": 523.1, "completions/max_terminated_length": 523.1, "completions/mean_length": 206.90130615234375, "completions/mean_terminated_length": 206.90130615234375, "completions/min_length": 48.5, "completions/min_terminated_length": 48.5, "entropy": 0.029879526576648156, "epoch": 5.342822161003979, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.026611328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 894247066.0, "reward": 8.129901266098022, "reward_std": 0.43008137345314024, "rewards/combined_reward/mean": 8.129901027679443, "rewards/combined_reward/std": 2.2682018160820006, "step": 8730 }, { "clip_ratio/high_max": 0.0020447293204294207, "clip_ratio/high_mean": 0.0011231948623996383, "clip_ratio/low_mean": 0.0008719516476655069, "clip_ratio/low_min": 0.00023549254134801837, "clip_ratio/region_mean": 0.0019951464962408255, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 194.67969360351563, "completions/mean_terminated_length": 194.67969360351563, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.029564369407792886, "epoch": 5.3611876339149065, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 897294796.0, "reward": 8.201043939590454, "reward_std": 0.42518138885498047, "rewards/combined_reward/mean": 8.20104341506958, "rewards/combined_reward/std": 2.1775040507316588, "step": 8760 }, { "clip_ratio/high_max": 0.0021585535374470054, "clip_ratio/high_mean": 0.0011840953513456043, "clip_ratio/low_mean": 0.0007683427360461792, "clip_ratio/low_min": 0.00016383212083989445, "clip_ratio/region_mean": 0.001952438101094837, "completions/clipped_ratio": 0.0, "completions/max_length": 543.8, "completions/max_terminated_length": 543.8, "completions/mean_length": 201.5135482788086, "completions/mean_terminated_length": 201.5135482788086, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.030259422585368158, "epoch": 5.379553106825834, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 900368768.0, "reward": 7.911883020401001, "reward_std": 0.37964055836200716, "rewards/combined_reward/mean": 7.91188268661499, "rewards/combined_reward/std": 2.281502628326416, "step": 8790 }, { "clip_ratio/high_max": 0.0024794448315030118, "clip_ratio/high_mean": 0.0013730999554657805, "clip_ratio/low_mean": 0.0008355331373119649, "clip_ratio/low_min": 0.00025081409694394097, "clip_ratio/region_mean": 0.0022086331053287724, "completions/clipped_ratio": 0.0, "completions/max_length": 561.5, "completions/max_terminated_length": 561.5, "completions/mean_length": 201.7018264770508, "completions/mean_terminated_length": 201.7018264770508, "completions/min_length": 46.1, "completions/min_terminated_length": 46.1, "entropy": 0.03336102861600618, "epoch": 5.3979185797367615, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 903443463.0, "reward": 8.153110694885253, "reward_std": 0.3186938151717186, "rewards/combined_reward/mean": 8.153110551834107, "rewards/combined_reward/std": 2.2316856026649474, "step": 8820 }, { "clip_ratio/high_max": 0.0018987872147893843, "clip_ratio/high_mean": 0.0010867244978726375, "clip_ratio/low_mean": 0.000721882819198072, "clip_ratio/low_min": 0.00022954339923065466, "clip_ratio/region_mean": 0.0018086073362307312, "completions/clipped_ratio": 0.0, "completions/max_length": 674.8, "completions/max_terminated_length": 674.8, "completions/mean_length": 205.39792175292968, "completions/mean_terminated_length": 205.39792175292968, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.030015433269242445, "epoch": 5.416284052647689, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 906532351.0, "reward": 8.063270425796508, "reward_std": 0.34909532517194747, "rewards/combined_reward/mean": 8.063270235061646, "rewards/combined_reward/std": 2.3172573924064634, "step": 8850 }, { "clip_ratio/high_max": 0.0020144608555710875, "clip_ratio/high_mean": 0.001197773959453722, "clip_ratio/low_mean": 0.0007713654722465435, "clip_ratio/low_min": 0.00026370647974545137, "clip_ratio/region_mean": 0.0019691394219989887, "completions/clipped_ratio": 0.0, "completions/max_length": 580.9, "completions/max_terminated_length": 580.9, "completions/mean_length": 203.14896240234376, "completions/mean_terminated_length": 203.14896240234376, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.034041551795477666, "epoch": 5.4346495255586165, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.007354736328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 909612603.0, "reward": 8.175152826309205, "reward_std": 0.31543236821889875, "rewards/combined_reward/mean": 8.175152349472047, "rewards/combined_reward/std": 2.4091323614120483, "step": 8880 }, { "clip_ratio/high_max": 0.001828701600121955, "clip_ratio/high_mean": 0.0009885817367224566, "clip_ratio/low_mean": 0.0007888891062369415, "clip_ratio/low_min": 0.00021161291539707842, "clip_ratio/region_mean": 0.0017774708210102593, "completions/clipped_ratio": 0.0, "completions/max_length": 654.6, "completions/max_terminated_length": 654.6, "completions/mean_length": 196.49167175292968, "completions/mean_terminated_length": 196.49167175292968, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.027988343220204114, "epoch": 5.453014998469544, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.005767822265625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 912667291.0, "reward": 8.00486330986023, "reward_std": 0.3665647774934769, "rewards/combined_reward/mean": 8.004863119125366, "rewards/combined_reward/std": 2.256450879573822, "step": 8910 }, { "clip_ratio/high_max": 0.0018711368730388736, "clip_ratio/high_mean": 0.001000276708509773, "clip_ratio/low_mean": 0.0008174954937809768, "clip_ratio/low_min": 0.00026033567555714396, "clip_ratio/region_mean": 0.0018177721974401114, "completions/clipped_ratio": 0.0, "completions/max_length": 661.2, "completions/max_terminated_length": 661.2, "completions/mean_length": 210.6916702270508, "completions/mean_terminated_length": 210.6916702270508, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.030685379868373275, "epoch": 5.4713804713804715, "frac_reward_zero_std": 0.14166666902601718, "grad_norm": 0.0244140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 915776507.0, "reward": 7.884799909591675, "reward_std": 0.3844747841358185, "rewards/combined_reward/mean": 7.884799480438232, "rewards/combined_reward/std": 2.380043613910675, "step": 8940 }, { "clip_ratio/high_max": 0.0018193690804764629, "clip_ratio/high_mean": 0.0009454541124190049, "clip_ratio/low_mean": 0.00075557081787944, "clip_ratio/low_min": 0.00017751267563047198, "clip_ratio/region_mean": 0.0017010249148976678, "completions/clipped_ratio": 0.0, "completions/max_length": 573.9, "completions/max_terminated_length": 573.9, "completions/mean_length": 195.29792022705078, "completions/mean_terminated_length": 195.29792022705078, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02696556826122105, "epoch": 5.489745944291399, "frac_reward_zero_std": 0.20000000447034835, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 918826611.0, "reward": 7.986700582504272, "reward_std": 0.3276138260960579, "rewards/combined_reward/mean": 7.98670015335083, "rewards/combined_reward/std": 2.477652931213379, "step": 8970 }, { "clip_ratio/high_max": 0.0020950568180220825, "clip_ratio/high_mean": 0.00123314863461322, "clip_ratio/low_mean": 0.0008167510385343727, "clip_ratio/low_min": 0.0002635040262248367, "clip_ratio/region_mean": 0.0020498996765430395, "completions/clipped_ratio": 0.0, "completions/max_length": 550.6, "completions/max_terminated_length": 550.6, "completions/mean_length": 198.5846389770508, "completions/mean_terminated_length": 198.5846389770508, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.029828370455652474, "epoch": 5.508111417202326, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.01068115234375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 921889336.0, "reward": 8.137866926193237, "reward_std": 0.3832588613033295, "rewards/combined_reward/mean": 8.137866640090943, "rewards/combined_reward/std": 2.0567649602890015, "step": 9000 }, { "clip_ratio/high_max": 0.002219511258590501, "clip_ratio/high_mean": 0.0012206477040308528, "clip_ratio/low_mean": 0.0008399822497570616, "clip_ratio/low_min": 0.00019210658501833677, "clip_ratio/region_mean": 0.002060629957607792, "completions/clipped_ratio": 0.0, "completions/max_length": 621.6, "completions/max_terminated_length": 621.6, "completions/mean_length": 200.33828430175782, "completions/mean_terminated_length": 200.33828430175782, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.031227933056652547, "epoch": 5.526476890113254, "frac_reward_zero_std": 0.1791666705161333, "grad_norm": 0.029296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 924958795.0, "reward": 7.883407211303711, "reward_std": 0.4289989709854126, "rewards/combined_reward/mean": 7.883406925201416, "rewards/combined_reward/std": 2.4343653678894044, "step": 9030 }, { "clip_ratio/high_max": 0.0018942307821513775, "clip_ratio/high_mean": 0.0011308166363354151, "clip_ratio/low_mean": 0.0007122277036614832, "clip_ratio/low_min": 0.00016770799241688412, "clip_ratio/region_mean": 0.001843044334964361, "completions/clipped_ratio": 0.0, "completions/max_length": 584.9, "completions/max_terminated_length": 584.9, "completions/mean_length": 194.56927490234375, "completions/mean_terminated_length": 194.56927490234375, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02628216869197786, "epoch": 5.544842363024181, "frac_reward_zero_std": 0.17083333432674408, "grad_norm": 0.02490234375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 928006101.0, "reward": 7.953952741622925, "reward_std": 0.3796884432435036, "rewards/combined_reward/mean": 7.9539525508880615, "rewards/combined_reward/std": 2.4344014286994935, "step": 9060 }, { "clip_ratio/high_max": 0.0021013997839569734, "clip_ratio/high_mean": 0.0012957371269294526, "clip_ratio/low_mean": 0.0007550181209808216, "clip_ratio/low_min": 0.00018540579476393758, "clip_ratio/region_mean": 0.0020507552248697418, "completions/clipped_ratio": 0.0, "completions/max_length": 588.3, "completions/max_terminated_length": 588.3, "completions/mean_length": 209.1151092529297, "completions/mean_terminated_length": 209.1151092529297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.030330325880398353, "epoch": 5.563207835935109, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 931109263.0, "reward": 7.909821653366089, "reward_std": 0.3849013239145279, "rewards/combined_reward/mean": 7.909821557998657, "rewards/combined_reward/std": 2.299412262439728, "step": 9090 }, { "clip_ratio/high_max": 0.0025736330205594032, "clip_ratio/high_mean": 0.0014873038537189132, "clip_ratio/low_mean": 0.0007718036671576556, "clip_ratio/low_min": 0.00020550086652898852, "clip_ratio/region_mean": 0.00225910752293809, "completions/clipped_ratio": 0.0, "completions/max_length": 617.3, "completions/max_terminated_length": 617.3, "completions/mean_length": 207.45989990234375, "completions/mean_terminated_length": 207.45989990234375, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03766235443763435, "epoch": 5.581573308846036, "frac_reward_zero_std": 0.1291666693985462, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 934206069.0, "reward": 8.203493690490722, "reward_std": 0.3911198079586029, "rewards/combined_reward/mean": 8.203493547439574, "rewards/combined_reward/std": 2.1994381189346313, "step": 9120 }, { "clip_ratio/high_max": 0.0018540634807626096, "clip_ratio/high_mean": 0.001040726720627087, "clip_ratio/low_mean": 0.0008428799045456496, "clip_ratio/low_min": 0.00028715304603489736, "clip_ratio/region_mean": 0.0018836066160777894, "completions/clipped_ratio": 0.0, "completions/max_length": 541.9, "completions/max_terminated_length": 541.9, "completions/mean_length": 209.7252670288086, "completions/mean_terminated_length": 209.7252670288086, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.028145201628406842, "epoch": 5.599938781756964, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 937311574.0, "reward": 8.111738920211792, "reward_std": 0.36892626583576205, "rewards/combined_reward/mean": 8.11173849105835, "rewards/combined_reward/std": 2.1856593251228333, "step": 9150 }, { "clip_ratio/high_max": 0.0021947679328150117, "clip_ratio/high_mean": 0.0012855613148227954, "clip_ratio/low_mean": 0.0008424064838133442, "clip_ratio/low_min": 0.00025341164194590723, "clip_ratio/region_mean": 0.002127967816098438, "completions/clipped_ratio": 0.0, "completions/max_length": 550.2, "completions/max_terminated_length": 550.2, "completions/mean_length": 187.4002670288086, "completions/mean_terminated_length": 187.4002670288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03071384006179869, "epoch": 5.618304254667891, "frac_reward_zero_std": 0.24583334103226662, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 940331351.0, "reward": 7.777670764923096, "reward_std": 0.34516645669937135, "rewards/combined_reward/mean": 7.777670526504517, "rewards/combined_reward/std": 2.4981096744537354, "step": 9180 }, { "clip_ratio/high_max": 0.0020214903343003243, "clip_ratio/high_mean": 0.0011230670896717734, "clip_ratio/low_mean": 0.0007196781693589098, "clip_ratio/low_min": 0.0002024444508909558, "clip_ratio/region_mean": 0.0018427452790395666, "completions/clipped_ratio": 0.0, "completions/max_length": 512.4, "completions/max_terminated_length": 512.4, "completions/mean_length": 199.52708740234374, "completions/mean_terminated_length": 199.52708740234374, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.028281930725400646, "epoch": 5.636669727578818, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 943397695.0, "reward": 7.993898391723633, "reward_std": 0.3195725679397583, "rewards/combined_reward/mean": 7.993898200988769, "rewards/combined_reward/std": 2.2945847392082213, "step": 9210 }, { "clip_ratio/high_max": 0.0019726915716698082, "clip_ratio/high_mean": 0.0011588441469939426, "clip_ratio/low_mean": 0.0008057568479368153, "clip_ratio/low_min": 0.00025502623951373, "clip_ratio/region_mean": 0.00196460100026646, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 195.84662017822265, "completions/mean_terminated_length": 195.84662017822265, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.028681498075214525, "epoch": 5.6550352004897455, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.006988525390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 946449906.0, "reward": 8.0490318775177, "reward_std": 0.3642562747001648, "rewards/combined_reward/mean": 8.049031734466553, "rewards/combined_reward/std": 2.3589600682258607, "step": 9240 }, { "clip_ratio/high_max": 0.001838652807540105, "clip_ratio/high_mean": 0.0010260744287127939, "clip_ratio/low_mean": 0.0008326410123117966, "clip_ratio/low_min": 0.00022245378713705576, "clip_ratio/region_mean": 0.001858715439933197, "completions/clipped_ratio": 0.0, "completions/max_length": 581.4, "completions/max_terminated_length": 581.4, "completions/mean_length": 188.9627670288086, "completions/mean_terminated_length": 188.9627670288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.0285551896939675, "epoch": 5.673400673400673, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 949475683.0, "reward": 7.952059888839722, "reward_std": 0.31233124136924745, "rewards/combined_reward/mean": 7.952059602737426, "rewards/combined_reward/std": 2.3309791564941404, "step": 9270 }, { "clip_ratio/high_max": 0.0021053641340889346, "clip_ratio/high_mean": 0.0011885932607886692, "clip_ratio/low_mean": 0.0008266603773639266, "clip_ratio/low_min": 0.00022422653152413355, "clip_ratio/region_mean": 0.0020152536604048994, "completions/clipped_ratio": 0.0, "completions/max_length": 585.8, "completions/max_terminated_length": 585.8, "completions/mean_length": 186.69636077880858, "completions/mean_terminated_length": 186.69636077880858, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.030737200969209275, "epoch": 5.6917661463116005, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 952492757.0, "reward": 8.014194774627686, "reward_std": 0.3343233987689018, "rewards/combined_reward/mean": 8.014194440841674, "rewards/combined_reward/std": 2.3008457005023955, "step": 9300 }, { "clip_ratio/high_max": 0.001985696624130166, "clip_ratio/high_mean": 0.0011819549043138977, "clip_ratio/low_mean": 0.0007956873632186519, "clip_ratio/low_min": 0.00023240761511260643, "clip_ratio/region_mean": 0.0019776422918463747, "completions/clipped_ratio": 0.0, "completions/max_length": 662.4, "completions/max_terminated_length": 662.4, "completions/mean_length": 206.5612030029297, "completions/mean_terminated_length": 206.5612030029297, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03249996619609495, "epoch": 5.710131619222528, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0289306640625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 955586112.0, "reward": 7.962494325637818, "reward_std": 0.40702575743198394, "rewards/combined_reward/mean": 7.962494421005249, "rewards/combined_reward/std": 2.464700400829315, "step": 9330 }, { "clip_ratio/high_max": 0.0016887425275247854, "clip_ratio/high_mean": 0.0009700113070721273, "clip_ratio/low_mean": 0.0006886558433810327, "clip_ratio/low_min": 0.000136429176927777, "clip_ratio/region_mean": 0.0016586671568802557, "completions/clipped_ratio": 0.0, "completions/max_length": 621.5, "completions/max_terminated_length": 621.5, "completions/mean_length": 194.8651092529297, "completions/mean_terminated_length": 194.8651092529297, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02701294912646214, "epoch": 5.7284970921334555, "frac_reward_zero_std": 0.12916667014360428, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 958634554.0, "reward": 8.05515947341919, "reward_std": 0.33827132135629656, "rewards/combined_reward/mean": 8.05515947341919, "rewards/combined_reward/std": 2.2798488259315492, "step": 9360 }, { "clip_ratio/high_max": 0.0019102429927443155, "clip_ratio/high_mean": 0.0010440698104503098, "clip_ratio/low_mean": 0.0007833859753494228, "clip_ratio/low_min": 0.00020956952284905127, "clip_ratio/region_mean": 0.0018274557767047857, "completions/clipped_ratio": 0.0, "completions/max_length": 517.6, "completions/max_terminated_length": 517.6, "completions/mean_length": 197.1320358276367, "completions/mean_terminated_length": 197.1320358276367, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.026957952906377612, "epoch": 5.746862565044383, "frac_reward_zero_std": 0.2291666731238365, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 961691701.0, "reward": 8.025496482849121, "reward_std": 0.3158004514873028, "rewards/combined_reward/mean": 8.025496244430542, "rewards/combined_reward/std": 2.2044416666030884, "step": 9390 }, { "clip_ratio/high_max": 0.0022242500315769577, "clip_ratio/high_mean": 0.0012094831849632707, "clip_ratio/low_mean": 0.0007643171406622666, "clip_ratio/low_min": 0.0001770735878380947, "clip_ratio/region_mean": 0.001973800335933144, "completions/clipped_ratio": 0.0, "completions/max_length": 571.9, "completions/max_terminated_length": 571.9, "completions/mean_length": 197.5276107788086, "completions/mean_terminated_length": 197.5276107788086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03153802778882285, "epoch": 5.7652280379553105, "frac_reward_zero_std": 0.16666666753590106, "grad_norm": 0.0084228515625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 964750367.0, "reward": 8.204388332366943, "reward_std": 0.29786168187856676, "rewards/combined_reward/mean": 8.204387903213501, "rewards/combined_reward/std": 2.108634054660797, "step": 9420 }, { "clip_ratio/high_max": 0.0018492906844282213, "clip_ratio/high_mean": 0.0010523971788643394, "clip_ratio/low_mean": 0.0008602169867420647, "clip_ratio/low_min": 0.00018814472350641153, "clip_ratio/region_mean": 0.0019126141759140106, "completions/clipped_ratio": 0.0, "completions/max_length": 543.6, "completions/max_terminated_length": 543.6, "completions/mean_length": 189.94219207763672, "completions/mean_terminated_length": 189.94219207763672, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.02815769298467785, "epoch": 5.783593510866238, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 967779905.0, "reward": 8.304992246627808, "reward_std": 0.354947804659605, "rewards/combined_reward/mean": 8.30499210357666, "rewards/combined_reward/std": 2.1090090751647947, "step": 9450 }, { "clip_ratio/high_max": 0.0019896340352715926, "clip_ratio/high_mean": 0.0010847851840177706, "clip_ratio/low_mean": 0.0006912479064946335, "clip_ratio/low_min": 0.0001475898990368781, "clip_ratio/region_mean": 0.0017760330859649305, "completions/clipped_ratio": 0.0, "completions/max_length": 579.7, "completions/max_terminated_length": 579.7, "completions/mean_length": 205.25000457763673, "completions/mean_terminated_length": 205.25000457763673, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.02699039278862377, "epoch": 5.8019589837771655, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 970868225.0, "reward": 8.133332204818725, "reward_std": 0.33679748326539993, "rewards/combined_reward/mean": 8.133331918716431, "rewards/combined_reward/std": 2.2395811557769774, "step": 9480 }, { "clip_ratio/high_max": 0.0019127341649436858, "clip_ratio/high_mean": 0.001040039544144141, "clip_ratio/low_mean": 0.0006701911303631883, "clip_ratio/low_min": 9.564126182037095e-05, "clip_ratio/region_mean": 0.0017102306827534144, "completions/clipped_ratio": 0.0, "completions/max_length": 536.9, "completions/max_terminated_length": 536.9, "completions/mean_length": 202.16927185058594, "completions/mean_terminated_length": 202.16927185058594, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.029656136641278863, "epoch": 5.820324456688093, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 973944715.0, "reward": 8.137972259521485, "reward_std": 0.3587422460317612, "rewards/combined_reward/mean": 8.137971925735474, "rewards/combined_reward/std": 2.2075275659561155, "step": 9510 }, { "clip_ratio/high_max": 0.0026005912038575235, "clip_ratio/high_mean": 0.0014506876910066542, "clip_ratio/low_mean": 0.0009306170663573236, "clip_ratio/low_min": 0.0002839217590614377, "clip_ratio/region_mean": 0.002381304794107564, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 194.8854232788086, "completions/mean_terminated_length": 194.8854232788086, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.03448852999135852, "epoch": 5.8386899295990204, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 976993235.0, "reward": 7.876341342926025, "reward_std": 0.4203170269727707, "rewards/combined_reward/mean": 7.876341056823731, "rewards/combined_reward/std": 2.3789541244506838, "step": 9540 }, { "clip_ratio/high_max": 0.0017057970520302963, "clip_ratio/high_mean": 0.001088335290478426, "clip_ratio/low_mean": 0.000608854856727703, "clip_ratio/low_min": 9.136168446275406e-05, "clip_ratio/region_mean": 0.0016971901568467729, "completions/clipped_ratio": 0.0, "completions/max_length": 617.8, "completions/max_terminated_length": 617.8, "completions/mean_length": 193.05469207763673, "completions/mean_terminated_length": 193.05469207763673, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.027653698157519103, "epoch": 5.857055402509948, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.007659912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 980034725.0, "reward": 8.169245624542237, "reward_std": 0.33899392932653427, "rewards/combined_reward/mean": 8.169245147705078, "rewards/combined_reward/std": 2.1091018676757813, "step": 9570 }, { "clip_ratio/high_max": 0.002161990998138208, "clip_ratio/high_mean": 0.001199472526544317, "clip_ratio/low_mean": 0.0007716075543915698, "clip_ratio/low_min": 0.0002384536281169858, "clip_ratio/region_mean": 0.001971080062018397, "completions/clipped_ratio": 0.0, "completions/max_length": 674.3, "completions/max_terminated_length": 674.3, "completions/mean_length": 204.4117233276367, "completions/mean_terminated_length": 204.4117233276367, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.035600173349181814, "epoch": 5.875420875420875, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 983119826.0, "reward": 8.248746824264526, "reward_std": 0.4439978986978531, "rewards/combined_reward/mean": 8.248746633529663, "rewards/combined_reward/std": 2.0833922266960143, "step": 9600 }, { "clip_ratio/high_max": 0.0024521025023811188, "clip_ratio/high_mean": 0.0013696039449617578, "clip_ratio/low_mean": 0.000783884590297627, "clip_ratio/low_min": 0.0001790842065626445, "clip_ratio/region_mean": 0.0021534885517515553, "completions/clipped_ratio": 0.0, "completions/max_length": 559.3, "completions/max_terminated_length": 559.3, "completions/mean_length": 192.84844055175782, "completions/mean_terminated_length": 192.84844055175782, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.031406818935647604, "epoch": 5.893786348331803, "frac_reward_zero_std": 0.18333334065973758, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 986160524.0, "reward": 8.005151510238647, "reward_std": 0.39243160337209704, "rewards/combined_reward/mean": 8.005150890350341, "rewards/combined_reward/std": 2.3131621837615968, "step": 9630 }, { "clip_ratio/high_max": 0.0021266812953399496, "clip_ratio/high_mean": 0.00124989616330519, "clip_ratio/low_mean": 0.0007743009004722505, "clip_ratio/low_min": 0.00023738240221670518, "clip_ratio/region_mean": 0.0020241970681430153, "completions/clipped_ratio": 0.0, "completions/max_length": 591.1, "completions/max_terminated_length": 591.1, "completions/mean_length": 202.81094055175782, "completions/mean_terminated_length": 202.81094055175782, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03377421443971495, "epoch": 5.91215182124273, "frac_reward_zero_std": 0.21666667461395264, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 989239478.0, "reward": 8.144988536834717, "reward_std": 0.39533293098211286, "rewards/combined_reward/mean": 8.144988250732421, "rewards/combined_reward/std": 2.1968421936035156, "step": 9660 }, { "clip_ratio/high_max": 0.001861696890652335, "clip_ratio/high_mean": 0.0011141062595318848, "clip_ratio/low_mean": 0.0007256596175769422, "clip_ratio/low_min": 0.0002031955222870844, "clip_ratio/region_mean": 0.0018397658864463058, "completions/clipped_ratio": 0.0, "completions/max_length": 583.8, "completions/max_terminated_length": 583.8, "completions/mean_length": 190.5822967529297, "completions/mean_terminated_length": 190.5822967529297, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.030592395178973674, "epoch": 5.930517294153658, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.004425048828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 992271474.0, "reward": 7.7275793075561525, "reward_std": 0.4418093651533127, "rewards/combined_reward/mean": 7.727579069137573, "rewards/combined_reward/std": 2.687576723098755, "step": 9690 }, { "clip_ratio/high_max": 0.0018664595428466177, "clip_ratio/high_mean": 0.0010422896643528172, "clip_ratio/low_mean": 0.0007441388474641523, "clip_ratio/low_min": 0.0001384586185546747, "clip_ratio/region_mean": 0.0017864285066025332, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 191.44271240234374, "completions/mean_terminated_length": 191.44271240234374, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027753580377126733, "epoch": 5.948882767064585, "frac_reward_zero_std": 0.15000000260770321, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 995306774.0, "reward": 7.8856359958648685, "reward_std": 0.3157066717743874, "rewards/combined_reward/mean": 7.885635852813721, "rewards/combined_reward/std": 2.309214007854462, "step": 9720 }, { "clip_ratio/high_max": 0.0018794179176135609, "clip_ratio/high_mean": 0.0010903092098790998, "clip_ratio/low_mean": 0.0008414405700023053, "clip_ratio/low_min": 0.00024249195436520192, "clip_ratio/region_mean": 0.0019317497761221602, "completions/clipped_ratio": 0.0, "completions/max_length": 576.9, "completions/max_terminated_length": 576.9, "completions/mean_length": 204.39662017822266, "completions/mean_terminated_length": 204.39662017822266, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03180947306876381, "epoch": 5.967248239975513, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0050048828125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 998391817.0, "reward": 7.927709150314331, "reward_std": 0.2852894887328148, "rewards/combined_reward/mean": 7.9277090549469, "rewards/combined_reward/std": 2.3993886351585387, "step": 9750 }, { "clip_ratio/high_max": 0.001816621447505895, "clip_ratio/high_mean": 0.0010275948045697684, "clip_ratio/low_mean": 0.0007904896230077914, "clip_ratio/low_min": 0.00021541950676085738, "clip_ratio/region_mean": 0.001818084436429975, "completions/clipped_ratio": 0.0, "completions/max_length": 636.9, "completions/max_terminated_length": 636.9, "completions/mean_length": 197.6674545288086, "completions/mean_terminated_length": 197.6674545288086, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.029886440047994255, "epoch": 5.98561371288644, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.043212890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1001451020.0, "reward": 7.946551656723022, "reward_std": 0.31676683723926546, "rewards/combined_reward/mean": 7.946551370620727, "rewards/combined_reward/std": 2.427341902256012, "step": 9780 }, { "clip_ratio/high_max": 0.002048569036500727, "clip_ratio/high_mean": 0.0010913253471266753, "clip_ratio/low_mean": 0.0006543557995979934, "clip_ratio/low_min": 0.00012148273836813684, "clip_ratio/region_mean": 0.0017456811658091477, "completions/clipped_ratio": 0.0, "completions/max_length": 576.8181818181819, "completions/max_terminated_length": 576.8181818181819, "completions/mean_length": 201.76089477539062, "completions/mean_terminated_length": 201.76089477539062, "completions/min_length": 39.45454545454545, "completions/min_terminated_length": 39.45454545454545, "entropy": 0.02699903305619955, "epoch": 6.00428527701255, "frac_reward_zero_std": 0.1553030332381075, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1004833434.0, "reward": 8.091922543265603, "reward_std": 0.3243704993616451, "rewards/combined_reward/mean": 8.09192232652144, "rewards/combined_reward/std": 2.188958103006536, "step": 9810 }, { "clip_ratio/high_max": 0.002534487714729039, "clip_ratio/high_mean": 0.0013886496890336275, "clip_ratio/low_mean": 0.0010407551812628905, "clip_ratio/low_min": 0.00033076682787699, "clip_ratio/region_mean": 0.002429404872721837, "completions/clipped_ratio": 0.0, "completions/max_length": 561.3, "completions/max_terminated_length": 561.3, "completions/mean_length": 197.60208892822266, "completions/mean_terminated_length": 197.60208892822266, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.02966399726768335, "epoch": 6.022650749923478, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1007892386.0, "reward": 7.828920030593872, "reward_std": 0.32187981754541395, "rewards/combined_reward/mean": 7.828919887542725, "rewards/combined_reward/std": 2.297757875919342, "step": 9840 }, { "clip_ratio/high_max": 0.0026147156468747804, "clip_ratio/high_mean": 0.0015320714420037499, "clip_ratio/low_mean": 0.0009895928807964083, "clip_ratio/low_min": 0.00029418168051051905, "clip_ratio/region_mean": 0.0025216643194047113, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 207.99193115234374, "completions/mean_terminated_length": 207.99193115234374, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03271682256987939, "epoch": 6.041016222834405, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1010991235.0, "reward": 8.093193912506104, "reward_std": 0.35726049095392226, "rewards/combined_reward/mean": 8.093193674087525, "rewards/combined_reward/std": 2.3478534817695618, "step": 9870 }, { "clip_ratio/high_max": 0.0022270796025016656, "clip_ratio/high_mean": 0.0012325634525041096, "clip_ratio/low_mean": 0.0011359360611095325, "clip_ratio/low_min": 0.00031851600845887637, "clip_ratio/region_mean": 0.002368499532652398, "completions/clipped_ratio": 0.0, "completions/max_length": 560.3, "completions/max_terminated_length": 560.3, "completions/mean_length": 204.05547485351562, "completions/mean_terminated_length": 204.05547485351562, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.028436168422922493, "epoch": 6.059381695745332, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1014074968.0, "reward": 7.883557891845703, "reward_std": 0.3630142599344254, "rewards/combined_reward/mean": 7.883557653427124, "rewards/combined_reward/std": 2.5246961236000063, "step": 9900 }, { "clip_ratio/high_max": 0.0024113568128086627, "clip_ratio/high_mean": 0.0013631554817645035, "clip_ratio/low_mean": 0.0011721279293851693, "clip_ratio/low_min": 0.00035908307448456376, "clip_ratio/region_mean": 0.002535283422912471, "completions/clipped_ratio": 0.0, "completions/max_length": 594.7, "completions/max_terminated_length": 594.7, "completions/mean_length": 207.02005767822266, "completions/mean_terminated_length": 207.02005767822266, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.0322588998824358, "epoch": 6.077747168656259, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.0084228515625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1017170085.0, "reward": 7.988095712661743, "reward_std": 0.2762982606887817, "rewards/combined_reward/mean": 7.988095474243164, "rewards/combined_reward/std": 2.332080352306366, "step": 9930 }, { "clip_ratio/high_max": 0.0021636920166201888, "clip_ratio/high_mean": 0.001145547021466579, "clip_ratio/low_mean": 0.000823029326784308, "clip_ratio/low_min": 0.00016469609799969477, "clip_ratio/region_mean": 0.001968576354071653, "completions/clipped_ratio": 0.0, "completions/max_length": 572.5, "completions/max_terminated_length": 572.5, "completions/mean_length": 196.41146545410157, "completions/mean_terminated_length": 196.41146545410157, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02651167979153494, "epoch": 6.096112641567187, "frac_reward_zero_std": 0.1708333410322666, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1020224465.0, "reward": 8.091043615341187, "reward_std": 0.3672377496957779, "rewards/combined_reward/mean": 8.091043472290039, "rewards/combined_reward/std": 2.337452566623688, "step": 9960 }, { "clip_ratio/high_max": 0.0024566232925280927, "clip_ratio/high_mean": 0.0013132236281914325, "clip_ratio/low_mean": 0.0010436563214170746, "clip_ratio/low_min": 0.0003248405788326636, "clip_ratio/region_mean": 0.002356879944272805, "completions/clipped_ratio": 0.0, "completions/max_length": 532.6, "completions/max_terminated_length": 532.6, "completions/mean_length": 199.3518295288086, "completions/mean_terminated_length": 199.3518295288086, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.028432809639101227, "epoch": 6.114478114478114, "frac_reward_zero_std": 0.16666667349636555, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1023290136.0, "reward": 8.207916498184204, "reward_std": 0.3539571762084961, "rewards/combined_reward/mean": 8.207916069030762, "rewards/combined_reward/std": 2.131142795085907, "step": 9990 }, { "clip_ratio/high_max": 0.002186179432707528, "clip_ratio/high_mean": 0.0013009280630891832, "clip_ratio/low_mean": 0.0010483744442656947, "clip_ratio/low_min": 0.0003581134131915557, "clip_ratio/region_mean": 0.002349302524817176, "completions/clipped_ratio": 0.0, "completions/max_length": 602.4, "completions/max_terminated_length": 602.4, "completions/mean_length": 214.17474365234375, "completions/mean_terminated_length": 214.17474365234375, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027797770973605414, "epoch": 6.132843587389042, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1026412727.0, "reward": 8.30265564918518, "reward_std": 0.3565302535891533, "rewards/combined_reward/mean": 8.302655124664307, "rewards/combined_reward/std": 2.002450907230377, "step": 10020 }, { "clip_ratio/high_max": 0.002699107078660745, "clip_ratio/high_mean": 0.0016104056094870126, "clip_ratio/low_mean": 0.0008314534256593713, "clip_ratio/low_min": 0.00021286839473759754, "clip_ratio/region_mean": 0.002441859026051437, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 187.5963577270508, "completions/mean_terminated_length": 187.5963577270508, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.032723352778702974, "epoch": 6.151209060299969, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.02294921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1029433257.0, "reward": 7.951911687850952, "reward_std": 0.3763223648071289, "rewards/combined_reward/mean": 7.951911401748657, "rewards/combined_reward/std": 2.231755816936493, "step": 10050 }, { "clip_ratio/high_max": 0.0025493705407522308, "clip_ratio/high_mean": 0.0014977970681987547, "clip_ratio/low_mean": 0.0010395007178885863, "clip_ratio/low_min": 0.0002333106846587422, "clip_ratio/region_mean": 0.002537297806702554, "completions/clipped_ratio": 0.0, "completions/max_length": 630.3, "completions/max_terminated_length": 630.3, "completions/mean_length": 198.8078170776367, "completions/mean_terminated_length": 198.8078170776367, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.033275303120414414, "epoch": 6.169574533210897, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1032496839.0, "reward": 7.873021411895752, "reward_std": 0.3700954347848892, "rewards/combined_reward/mean": 7.873021173477173, "rewards/combined_reward/std": 2.496058130264282, "step": 10080 }, { "clip_ratio/high_max": 0.0021592807352135424, "clip_ratio/high_mean": 0.001286079815023792, "clip_ratio/low_mean": 0.0008871068287893044, "clip_ratio/low_min": 0.00023266304172769498, "clip_ratio/region_mean": 0.002173186633444857, "completions/clipped_ratio": 0.0, "completions/max_length": 510.9, "completions/max_terminated_length": 510.9, "completions/mean_length": 187.00156860351564, "completions/mean_terminated_length": 187.00156860351564, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02731635581391553, "epoch": 6.187940006121824, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 1035515085.0, "reward": 7.845398330688477, "reward_std": 0.3214344337582588, "rewards/combined_reward/mean": 7.84539828300476, "rewards/combined_reward/std": 2.6498916029930113, "step": 10110 }, { "clip_ratio/high_max": 0.0027005000796634705, "clip_ratio/high_mean": 0.0015009555905029022, "clip_ratio/low_mean": 0.000871931845419264, "clip_ratio/low_min": 0.0002319292922038585, "clip_ratio/region_mean": 0.0023728874415004006, "completions/clipped_ratio": 0.0, "completions/max_length": 571.7, "completions/max_terminated_length": 571.7, "completions/mean_length": 198.0510467529297, "completions/mean_terminated_length": 198.0510467529297, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.028841399184117714, "epoch": 6.206305479032752, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.01275634765625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1038575761.0, "reward": 8.06613187789917, "reward_std": 0.3988070607185364, "rewards/combined_reward/mean": 8.066131353378296, "rewards/combined_reward/std": 2.262224519252777, "step": 10140 }, { "clip_ratio/high_max": 0.002662142528182206, "clip_ratio/high_mean": 0.0016028118203394114, "clip_ratio/low_mean": 0.0010221320420896519, "clip_ratio/low_min": 0.0003235206177729803, "clip_ratio/region_mean": 0.002624943849029175, "completions/clipped_ratio": 0.0, "completions/max_length": 504.9, "completions/max_terminated_length": 504.9, "completions/mean_length": 200.8401092529297, "completions/mean_terminated_length": 200.8401092529297, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.030117217265069485, "epoch": 6.224670951943679, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1041647147.0, "reward": 7.969760799407959, "reward_std": 0.2759613558650017, "rewards/combined_reward/mean": 7.96976056098938, "rewards/combined_reward/std": 2.3958715200424194, "step": 10170 }, { "clip_ratio/high_max": 0.0024974784561588117, "clip_ratio/high_mean": 0.00152488484151642, "clip_ratio/low_mean": 0.0011643243712872694, "clip_ratio/low_min": 0.0003555495665447476, "clip_ratio/region_mean": 0.002689209198191141, "completions/clipped_ratio": 0.0, "completions/max_length": 638.2, "completions/max_terminated_length": 638.2, "completions/mean_length": 200.6018280029297, "completions/mean_terminated_length": 200.6018280029297, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.031976124209662275, "epoch": 6.243036424854607, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1044717618.0, "reward": 8.012427282333373, "reward_std": 0.39158132672309875, "rewards/combined_reward/mean": 8.012427234649659, "rewards/combined_reward/std": 2.336369574069977, "step": 10200 }, { "clip_ratio/high_max": 0.0022634003209532237, "clip_ratio/high_mean": 0.0013145636151117893, "clip_ratio/low_mean": 0.0010569104969666418, "clip_ratio/low_min": 0.00027544721500210776, "clip_ratio/region_mean": 0.002371474085763718, "completions/clipped_ratio": 0.0, "completions/max_length": 590.1, "completions/max_terminated_length": 590.1, "completions/mean_length": 197.5862030029297, "completions/mean_terminated_length": 197.5862030029297, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.028690845736612878, "epoch": 6.261401897765534, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1047776509.0, "reward": 8.2620698928833, "reward_std": 0.35310458913445475, "rewards/combined_reward/mean": 8.262069702148438, "rewards/combined_reward/std": 1.9631313920021056, "step": 10230 }, { "clip_ratio/high_max": 0.002219181479207085, "clip_ratio/high_mean": 0.0012913724788935117, "clip_ratio/low_mean": 0.0009538947779219598, "clip_ratio/low_min": 0.00024644357205640214, "clip_ratio/region_mean": 0.0022452672526317957, "completions/clipped_ratio": 0.0, "completions/max_length": 581.3, "completions/max_terminated_length": 581.3, "completions/mean_length": 201.01172637939453, "completions/mean_terminated_length": 201.01172637939453, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.02889426308684051, "epoch": 6.279767370676462, "frac_reward_zero_std": 0.14166666865348815, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1050848554.0, "reward": 8.116506195068359, "reward_std": 0.37198122441768644, "rewards/combined_reward/mean": 8.116506004333496, "rewards/combined_reward/std": 2.249260425567627, "step": 10260 }, { "clip_ratio/high_max": 0.00251962029239318, "clip_ratio/high_mean": 0.0013582012534849733, "clip_ratio/low_mean": 0.00095061066482837, "clip_ratio/low_min": 0.00029243615426821634, "clip_ratio/region_mean": 0.002308811925953099, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 195.6541732788086, "completions/mean_terminated_length": 195.6541732788086, "completions/min_length": 46.1, "completions/min_terminated_length": 46.1, "entropy": 0.030602628458291294, "epoch": 6.298132843587389, "frac_reward_zero_std": 0.2000000037252903, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0031, "num_tokens": 1053900026.0, "reward": 8.36789960861206, "reward_std": 0.4036068424582481, "rewards/combined_reward/mean": 8.36789927482605, "rewards/combined_reward/std": 1.8821494340896607, "step": 10290 }, { "clip_ratio/high_max": 0.0021673680244324108, "clip_ratio/high_mean": 0.0011876317933153283, "clip_ratio/low_mean": 0.001039154092783671, "clip_ratio/low_min": 0.00026188463840905266, "clip_ratio/region_mean": 0.002226785900226484, "completions/clipped_ratio": 0.0, "completions/max_length": 657.4, "completions/max_terminated_length": 657.4, "completions/mean_length": 201.96485137939453, "completions/mean_terminated_length": 201.96485137939453, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03143616447535654, "epoch": 6.316498316498317, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.03173828125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 1056975731.0, "reward": 7.900217485427857, "reward_std": 0.38181168735027315, "rewards/combined_reward/mean": 7.9002172470092775, "rewards/combined_reward/std": 2.431005573272705, "step": 10320 }, { "clip_ratio/high_max": 0.0028342022487777285, "clip_ratio/high_mean": 0.001432271946517479, "clip_ratio/low_mean": 0.0009261413240892579, "clip_ratio/low_min": 0.0002546117842333236, "clip_ratio/region_mean": 0.002358413260662928, "completions/clipped_ratio": 0.0, "completions/max_length": 534.8, "completions/max_terminated_length": 534.8, "completions/mean_length": 203.56354675292968, "completions/mean_terminated_length": 203.56354675292968, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028806156680608788, "epoch": 6.334863789409244, "frac_reward_zero_std": 0.1291666693985462, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1060057575.0, "reward": 7.881789016723633, "reward_std": 0.32157984673976897, "rewards/combined_reward/mean": 7.881788539886474, "rewards/combined_reward/std": 2.532662093639374, "step": 10350 }, { "clip_ratio/high_max": 0.00259786124661332, "clip_ratio/high_mean": 0.0015198840245526904, "clip_ratio/low_mean": 0.0008848546314159951, "clip_ratio/low_min": 0.00022153296085889452, "clip_ratio/region_mean": 0.0024047386492990577, "completions/clipped_ratio": 0.0, "completions/max_length": 643.8, "completions/max_terminated_length": 643.8, "completions/mean_length": 199.62708892822266, "completions/mean_terminated_length": 199.62708892822266, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.028382741830622157, "epoch": 6.353229262320172, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.006256103515625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1063124303.0, "reward": 8.053720474243164, "reward_std": 0.37283010482788087, "rewards/combined_reward/mean": 8.053720283508301, "rewards/combined_reward/std": 2.316775417327881, "step": 10380 }, { "clip_ratio/high_max": 0.0024746413924731313, "clip_ratio/high_mean": 0.0012457209319109096, "clip_ratio/low_mean": 0.000546799115545582, "clip_ratio/low_min": 9.928959116223268e-05, "clip_ratio/region_mean": 0.0017925200430909173, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 204.60156631469727, "completions/mean_terminated_length": 204.60156631469727, "completions/min_length": 42.75, "completions/min_terminated_length": 42.75, "entropy": 0.028258833568543196, "epoch": 6.370982552800735, "frac_reward_zero_std": 0.13541666977107525, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1066192085.0, "reward": 8.139312267303467, "reward_std": 0.45484400540590286, "rewards/combined_reward/mean": 8.139312148094177, "rewards/combined_reward/std": 2.0818012356758118, "step": 10410 }, { "clip_ratio/high_max": 0.0021046127784453954, "clip_ratio/high_mean": 0.0011747550289631667, "clip_ratio/low_mean": 0.0008499215022917876, "clip_ratio/low_min": 0.00022767367724251623, "clip_ratio/region_mean": 0.0020246765207654486, "completions/clipped_ratio": 0.0, "completions/max_length": 545.5, "completions/max_terminated_length": 545.5, "completions/mean_length": 203.54922180175782, "completions/mean_terminated_length": 203.54922180175782, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.033023057303701836, "epoch": 6.389348025711662, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1069273874.0, "reward": 7.903672456741333, "reward_std": 0.36519409865140917, "rewards/combined_reward/mean": 7.903672218322754, "rewards/combined_reward/std": 2.409601056575775, "step": 10440 }, { "clip_ratio/high_max": 0.0018085472179033483, "clip_ratio/high_mean": 0.0010640010054582187, "clip_ratio/low_mean": 0.0008502026009712912, "clip_ratio/low_min": 0.00028845965207437987, "clip_ratio/region_mean": 0.0019142036049743182, "completions/clipped_ratio": 0.0, "completions/max_length": 554.6, "completions/max_terminated_length": 554.6, "completions/mean_length": 205.10391082763672, "completions/mean_terminated_length": 205.10391082763672, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.029624029714614152, "epoch": 6.40771349862259, "frac_reward_zero_std": 0.15833333767950536, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1072361633.0, "reward": 8.147389936447144, "reward_std": 0.2914060652256012, "rewards/combined_reward/mean": 8.1473895072937, "rewards/combined_reward/std": 2.1825822472572325, "step": 10470 }, { "clip_ratio/high_max": 0.0019840066403655023, "clip_ratio/high_mean": 0.0010962066817228333, "clip_ratio/low_mean": 0.0007288556739392031, "clip_ratio/low_min": 0.00017476283295157676, "clip_ratio/region_mean": 0.0018250623574810258, "completions/clipped_ratio": 0.0, "completions/max_length": 672.6, "completions/max_terminated_length": 672.6, "completions/mean_length": 203.56016235351564, "completions/mean_terminated_length": 203.56016235351564, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.03176993207695584, "epoch": 6.426078971533517, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.025146484375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1075443464.0, "reward": 8.170088911056519, "reward_std": 0.3729468435049057, "rewards/combined_reward/mean": 8.17008867263794, "rewards/combined_reward/std": 2.355090248584747, "step": 10500 }, { "clip_ratio/high_max": 0.00215026563819265, "clip_ratio/high_mean": 0.001231664010992972, "clip_ratio/low_mean": 0.0007552938324200416, "clip_ratio/low_min": 0.00019910361734218896, "clip_ratio/region_mean": 0.0019869578536599874, "completions/clipped_ratio": 0.0, "completions/max_length": 589.1, "completions/max_terminated_length": 589.1, "completions/mean_length": 198.47865142822266, "completions/mean_terminated_length": 198.47865142822266, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.03240077515753607, "epoch": 6.444444444444445, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1078505782.0, "reward": 7.863916492462158, "reward_std": 0.42669802755117414, "rewards/combined_reward/mean": 7.863916063308716, "rewards/combined_reward/std": 2.528838348388672, "step": 10530 }, { "clip_ratio/high_max": 0.0018984280235599727, "clip_ratio/high_mean": 0.001060227874404518, "clip_ratio/low_mean": 0.0008668418922752608, "clip_ratio/low_min": 0.00030585295341249247, "clip_ratio/region_mean": 0.0019270697676499063, "completions/clipped_ratio": 0.0, "completions/max_length": 667.9, "completions/max_terminated_length": 667.9, "completions/mean_length": 211.60625457763672, "completions/mean_terminated_length": 211.60625457763672, "completions/min_length": 50.4, "completions/min_terminated_length": 50.4, "entropy": 0.02839171759939442, "epoch": 6.462809917355372, "frac_reward_zero_std": 0.1416666679084301, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1081618510.0, "reward": 8.047037029266358, "reward_std": 0.33311879485845564, "rewards/combined_reward/mean": 8.047036981582641, "rewards/combined_reward/std": 2.334270882606506, "step": 10560 }, { "clip_ratio/high_max": 0.0019453762613314515, "clip_ratio/high_mean": 0.0010630559578809577, "clip_ratio/low_mean": 0.0007329613942905174, "clip_ratio/low_min": 0.0002084114123135805, "clip_ratio/region_mean": 0.001796017349503624, "completions/clipped_ratio": 0.0, "completions/max_length": 632.3, "completions/max_terminated_length": 632.3, "completions/mean_length": 198.52500610351564, "completions/mean_terminated_length": 198.52500610351564, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029496578422064584, "epoch": 6.4811753902663, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1084681006.0, "reward": 7.927882814407349, "reward_std": 0.41967497766017914, "rewards/combined_reward/mean": 7.927882671356201, "rewards/combined_reward/std": 2.385457956790924, "step": 10590 }, { "clip_ratio/high_max": 0.002006408544063258, "clip_ratio/high_mean": 0.001128528205542049, "clip_ratio/low_mean": 0.0008088822068505881, "clip_ratio/low_min": 0.00020825798346777447, "clip_ratio/region_mean": 0.0019374104371915261, "completions/clipped_ratio": 0.0, "completions/max_length": 523.6, "completions/max_terminated_length": 523.6, "completions/mean_length": 199.69323272705077, "completions/mean_terminated_length": 199.69323272705077, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.029849089806278546, "epoch": 6.499540863177227, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.024658203125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1087747988.0, "reward": 8.166810846328735, "reward_std": 0.32323179841041566, "rewards/combined_reward/mean": 8.166810655593872, "rewards/combined_reward/std": 2.067026102542877, "step": 10620 }, { "clip_ratio/high_max": 0.0017238872984307817, "clip_ratio/high_mean": 0.0009765882020777402, "clip_ratio/low_mean": 0.0007110147975860552, "clip_ratio/low_min": 0.00019614801446247535, "clip_ratio/region_mean": 0.0016876029942068272, "completions/clipped_ratio": 0.0, "completions/max_length": 667.2, "completions/max_terminated_length": 667.2, "completions/mean_length": 198.60599365234376, "completions/mean_terminated_length": 198.60599365234376, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.027284875232726335, "epoch": 6.517906336088155, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.02783203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1090810795.0, "reward": 8.070860528945923, "reward_std": 0.4024370551109314, "rewards/combined_reward/mean": 8.07086009979248, "rewards/combined_reward/std": 2.2022512555122375, "step": 10650 }, { "clip_ratio/high_max": 0.0018865249105147087, "clip_ratio/high_mean": 0.0010000663559670405, "clip_ratio/low_mean": 0.0006802355139977105, "clip_ratio/low_min": 0.00023739675792360987, "clip_ratio/region_mean": 0.0016803018569286601, "completions/clipped_ratio": 0.0, "completions/max_length": 622.4, "completions/max_terminated_length": 622.4, "completions/mean_length": 197.55208587646484, "completions/mean_terminated_length": 197.55208587646484, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.028688455543791253, "epoch": 6.536271808999082, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1093869555.0, "reward": 7.947677278518677, "reward_std": 0.3781751647591591, "rewards/combined_reward/mean": 7.947677183151245, "rewards/combined_reward/std": 2.4896944761276245, "step": 10680 }, { "clip_ratio/high_max": 0.001968346991634462, "clip_ratio/high_mean": 0.00109683546446225, "clip_ratio/low_mean": 0.0007246414766996168, "clip_ratio/low_min": 0.00021190736782349026, "clip_ratio/region_mean": 0.0018214769375238878, "completions/clipped_ratio": 0.0, "completions/max_length": 580.2, "completions/max_terminated_length": 580.2, "completions/mean_length": 198.94245147705078, "completions/mean_terminated_length": 198.94245147705078, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.028051188541576268, "epoch": 6.55463728191001, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1096933654.0, "reward": 7.908496332168579, "reward_std": 0.370561920106411, "rewards/combined_reward/mean": 7.908496141433716, "rewards/combined_reward/std": 2.3325585246086122, "step": 10710 }, { "clip_ratio/high_max": 0.0023871912688870605, "clip_ratio/high_mean": 0.0014116416835652974, "clip_ratio/low_mean": 0.000882115480756814, "clip_ratio/low_min": 0.0003038574124123746, "clip_ratio/region_mean": 0.0022937571544995685, "completions/clipped_ratio": 0.0, "completions/max_length": 646.5, "completions/max_terminated_length": 646.5, "completions/mean_length": 215.72578887939454, "completions/mean_terminated_length": 215.72578887939454, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.034846737033997974, "epoch": 6.573002754820937, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.0255126953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1100062201.0, "reward": 8.073169136047364, "reward_std": 0.3577323243021965, "rewards/combined_reward/mean": 8.073168802261353, "rewards/combined_reward/std": 2.211756134033203, "step": 10740 }, { "clip_ratio/high_max": 0.0023691792409711826, "clip_ratio/high_mean": 0.0014006383730399345, "clip_ratio/low_mean": 0.0009076942053070525, "clip_ratio/low_min": 0.00026912139292107896, "clip_ratio/region_mean": 0.0023083325731325507, "completions/clipped_ratio": 0.0, "completions/max_length": 598.6, "completions/max_terminated_length": 598.6, "completions/mean_length": 204.70156707763672, "completions/mean_terminated_length": 204.70156707763672, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.034669721700871986, "epoch": 6.591368227731865, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1103148415.0, "reward": 8.112948656082153, "reward_std": 0.3622673094272614, "rewards/combined_reward/mean": 8.11294846534729, "rewards/combined_reward/std": 2.2661693930625915, "step": 10770 }, { "clip_ratio/high_max": 0.002286372224140602, "clip_ratio/high_mean": 0.0013347121785045601, "clip_ratio/low_mean": 0.0009786538873110355, "clip_ratio/low_min": 0.0003133135765286473, "clip_ratio/region_mean": 0.002313366058903436, "completions/clipped_ratio": 0.0, "completions/max_length": 671.1, "completions/max_terminated_length": 671.1, "completions/mean_length": 202.2721405029297, "completions/mean_terminated_length": 202.2721405029297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.033837089252968626, "epoch": 6.609733700642791, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1106225300.0, "reward": 8.085478830337525, "reward_std": 0.3332758158445358, "rewards/combined_reward/mean": 8.085478591918946, "rewards/combined_reward/std": 2.2438681602478026, "step": 10800 }, { "clip_ratio/high_max": 0.0019107282976619899, "clip_ratio/high_mean": 0.0010871536155415622, "clip_ratio/low_mean": 0.000757499369319703, "clip_ratio/low_min": 0.00025252628159554055, "clip_ratio/region_mean": 0.0018446529837092384, "completions/clipped_ratio": 0.0, "completions/max_length": 511.3, "completions/max_terminated_length": 511.3, "completions/mean_length": 193.39167327880858, "completions/mean_terminated_length": 193.39167327880858, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.028212145001937947, "epoch": 6.628099173553719, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1109268084.0, "reward": 7.84516453742981, "reward_std": 0.34574130177497864, "rewards/combined_reward/mean": 7.845164442062378, "rewards/combined_reward/std": 2.354579758644104, "step": 10830 }, { "clip_ratio/high_max": 0.0020044857592438348, "clip_ratio/high_mean": 0.0011511855964878729, "clip_ratio/low_mean": 0.0008802816359093413, "clip_ratio/low_min": 0.0002557177171790196, "clip_ratio/region_mean": 0.0020314672568929383, "completions/clipped_ratio": 0.0, "completions/max_length": 534.9, "completions/max_terminated_length": 534.9, "completions/mean_length": 193.02292022705078, "completions/mean_terminated_length": 193.02292022705078, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029017647736084957, "epoch": 6.646464646464646, "frac_reward_zero_std": 0.14583333544433116, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1112309452.0, "reward": 7.99456033706665, "reward_std": 0.36289649158716203, "rewards/combined_reward/mean": 7.99456000328064, "rewards/combined_reward/std": 2.311826193332672, "step": 10860 }, { "clip_ratio/high_max": 0.0017915730973375806, "clip_ratio/high_mean": 0.0010788316949401633, "clip_ratio/low_mean": 0.0007780123226742338, "clip_ratio/low_min": 0.0002444269511518845, "clip_ratio/region_mean": 0.0018568440097927426, "completions/clipped_ratio": 0.0, "completions/max_length": 772.4, "completions/max_terminated_length": 772.4, "completions/mean_length": 204.38333892822266, "completions/mean_terminated_length": 204.38333892822266, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03103957246057689, "epoch": 6.664830119375574, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1115394444.0, "reward": 8.07150239944458, "reward_std": 0.34466769695281985, "rewards/combined_reward/mean": 8.071502208709717, "rewards/combined_reward/std": 2.3442142963409425, "step": 10890 }, { "clip_ratio/high_max": 0.0020444499285076747, "clip_ratio/high_mean": 0.0010179195674330306, "clip_ratio/low_mean": 0.0007100335883781857, "clip_ratio/low_min": 0.00015878050714187946, "clip_ratio/region_mean": 0.001727953165633759, "completions/clipped_ratio": 0.0, "completions/max_length": 558.5, "completions/max_terminated_length": 558.5, "completions/mean_length": 183.08568267822267, "completions/mean_terminated_length": 183.08568267822267, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.02786393459731092, "epoch": 6.683195592286501, "frac_reward_zero_std": 0.2250000059604645, "grad_norm": 0.00482177734375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1118397653.0, "reward": 8.04307188987732, "reward_std": 0.26040777042508123, "rewards/combined_reward/mean": 8.043071746826172, "rewards/combined_reward/std": 2.290791368484497, "step": 10920 }, { "clip_ratio/high_max": 0.0017270026944364266, "clip_ratio/high_mean": 0.000958368206435504, "clip_ratio/low_mean": 0.0007768953246947301, "clip_ratio/low_min": 0.00026335260602839604, "clip_ratio/region_mean": 0.0017352635453183514, "completions/clipped_ratio": 0.0, "completions/max_length": 545.3, "completions/max_terminated_length": 545.3, "completions/mean_length": 195.36953582763672, "completions/mean_terminated_length": 195.36953582763672, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.030484264545763532, "epoch": 6.701561065197429, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1121448032.0, "reward": 8.075290155410766, "reward_std": 0.32839497923851013, "rewards/combined_reward/mean": 8.075289916992187, "rewards/combined_reward/std": 2.3026638627052307, "step": 10950 }, { "clip_ratio/high_max": 0.0018794351715769153, "clip_ratio/high_mean": 0.0011947114344366128, "clip_ratio/low_mean": 0.0007277981467874876, "clip_ratio/low_min": 0.00021846320378244856, "clip_ratio/region_mean": 0.00192250957673726, "completions/clipped_ratio": 0.0, "completions/max_length": 670.4, "completions/max_terminated_length": 670.4, "completions/mean_length": 206.5773498535156, "completions/mean_terminated_length": 206.5773498535156, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03068149179841081, "epoch": 6.719926538108356, "frac_reward_zero_std": 0.13750000558793546, "grad_norm": 0.00372314453125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 1124541449.0, "reward": 7.898937654495239, "reward_std": 0.4102569378912449, "rewards/combined_reward/mean": 7.898937320709228, "rewards/combined_reward/std": 2.513041925430298, "step": 10980 }, { "clip_ratio/high_max": 0.001692190241010394, "clip_ratio/high_mean": 0.0009704587767676761, "clip_ratio/low_mean": 0.0007168147020517305, "clip_ratio/low_min": 0.00022554905275076937, "clip_ratio/region_mean": 0.0016872734709371191, "completions/clipped_ratio": 0.0, "completions/max_length": 618.6, "completions/max_terminated_length": 618.6, "completions/mean_length": 192.79427642822264, "completions/mean_terminated_length": 192.79427642822264, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.027644804632291197, "epoch": 6.738292011019284, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1127581939.0, "reward": 8.006838130950928, "reward_std": 0.27368214428424836, "rewards/combined_reward/mean": 8.006837844848633, "rewards/combined_reward/std": 2.3156854450702666, "step": 11010 }, { "clip_ratio/high_max": 0.0018573044731359308, "clip_ratio/high_mean": 0.0010273613222428442, "clip_ratio/low_mean": 0.0009104848590747376, "clip_ratio/low_min": 0.000319756433115496, "clip_ratio/region_mean": 0.0019378461787709967, "completions/clipped_ratio": 0.0, "completions/max_length": 678.2, "completions/max_terminated_length": 678.2, "completions/mean_length": 200.96042175292968, "completions/mean_terminated_length": 200.96042175292968, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.031826596893370154, "epoch": 6.756657483930211, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1130653787.0, "reward": 8.156072282791138, "reward_std": 0.30246098041534425, "rewards/combined_reward/mean": 8.156071758270263, "rewards/combined_reward/std": 2.102001613378525, "step": 11040 }, { "clip_ratio/high_max": 0.0020208085853179606, "clip_ratio/high_mean": 0.001062889106591077, "clip_ratio/low_mean": 0.0006864115609156821, "clip_ratio/low_min": 0.00015397015449707395, "clip_ratio/region_mean": 0.0017493006695682805, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 194.7062545776367, "completions/mean_terminated_length": 194.7062545776367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027308088587597014, "epoch": 6.775022956841139, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1133701619.0, "reward": 8.295580959320068, "reward_std": 0.33292365074157715, "rewards/combined_reward/mean": 8.295580673217774, "rewards/combined_reward/std": 1.983592116832733, "step": 11070 }, { "clip_ratio/high_max": 0.0016368008643136516, "clip_ratio/high_mean": 0.0009899170610879083, "clip_ratio/low_mean": 0.0007153577854599765, "clip_ratio/low_min": 0.0001666580067346028, "clip_ratio/region_mean": 0.0017052748439406666, "completions/clipped_ratio": 0.0, "completions/max_length": 598.4, "completions/max_terminated_length": 598.4, "completions/mean_length": 198.1166732788086, "completions/mean_terminated_length": 198.1166732788086, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.028851004938284557, "epoch": 6.793388429752066, "frac_reward_zero_std": 0.12500000447034837, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 1136762547.0, "reward": 8.05459017753601, "reward_std": 0.3888596177101135, "rewards/combined_reward/mean": 8.054590129852295, "rewards/combined_reward/std": 2.387381041049957, "step": 11100 }, { "clip_ratio/high_max": 0.0018630835341658287, "clip_ratio/high_mean": 0.0010847865499575466, "clip_ratio/low_mean": 0.0006589945519711667, "clip_ratio/low_min": 0.00016528503377533828, "clip_ratio/region_mean": 0.0017437811101747988, "completions/clipped_ratio": 0.0, "completions/max_length": 577.5, "completions/max_terminated_length": 577.5, "completions/mean_length": 204.50729675292968, "completions/mean_terminated_length": 204.50729675292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.027518632045636576, "epoch": 6.811753902662994, "frac_reward_zero_std": 0.19583334177732467, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1139848015.0, "reward": 8.306055402755737, "reward_std": 0.35786367058753965, "rewards/combined_reward/mean": 8.306054973602295, "rewards/combined_reward/std": 1.953884518146515, "step": 11130 }, { "clip_ratio/high_max": 0.0019269188914525634, "clip_ratio/high_mean": 0.0011062322123810493, "clip_ratio/low_mean": 0.0008858053416285354, "clip_ratio/low_min": 0.0002852065551754398, "clip_ratio/region_mean": 0.0019920375615280744, "completions/clipped_ratio": 0.0, "completions/max_length": 481.1, "completions/max_terminated_length": 481.1, "completions/mean_length": 194.34974365234376, "completions/mean_terminated_length": 194.34974365234376, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03205670472234488, "epoch": 6.830119375573921, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.024658203125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1142894478.0, "reward": 7.966527080535888, "reward_std": 0.36513111293315886, "rewards/combined_reward/mean": 7.966526889801026, "rewards/combined_reward/std": 2.3468013525009157, "step": 11160 }, { "clip_ratio/high_max": 0.001683691787426748, "clip_ratio/high_mean": 0.0010367452137870713, "clip_ratio/low_mean": 0.0007912226136492488, "clip_ratio/low_min": 0.00023447454077540897, "clip_ratio/region_mean": 0.0018279678134907348, "completions/clipped_ratio": 0.0, "completions/max_length": 682.1, "completions/max_terminated_length": 682.1, "completions/mean_length": 195.08047485351562, "completions/mean_terminated_length": 195.08047485351562, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02838400666757176, "epoch": 6.848484848484849, "frac_reward_zero_std": 0.22916667386889458, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1145943747.0, "reward": 8.121463394165039, "reward_std": 0.32643107771873475, "rewards/combined_reward/mean": 8.121463251113891, "rewards/combined_reward/std": 2.245337700843811, "step": 11190 }, { "clip_ratio/high_max": 0.002076488855527714, "clip_ratio/high_mean": 0.0011877023959338355, "clip_ratio/low_mean": 0.0007338394357551199, "clip_ratio/low_min": 0.0002102539231903696, "clip_ratio/region_mean": 0.0019215418142266572, "completions/clipped_ratio": 0.0, "completions/max_length": 624.4, "completions/max_terminated_length": 624.4, "completions/mean_length": 202.2697982788086, "completions/mean_terminated_length": 202.2697982788086, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.031160664558410645, "epoch": 6.866850321395776, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1149020623.0, "reward": 8.343217897415162, "reward_std": 0.365315680205822, "rewards/combined_reward/mean": 8.343217849731445, "rewards/combined_reward/std": 1.9464456677436828, "step": 11220 }, { "clip_ratio/high_max": 0.002265812174785727, "clip_ratio/high_mean": 0.0012815284953830997, "clip_ratio/low_mean": 0.0008367897181112009, "clip_ratio/low_min": 0.00023254654200475974, "clip_ratio/region_mean": 0.0021183182145856944, "completions/clipped_ratio": 0.0, "completions/max_length": 574.5, "completions/max_terminated_length": 574.5, "completions/mean_length": 189.24271392822266, "completions/mean_terminated_length": 189.24271392822266, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03272278867661953, "epoch": 6.885215794306704, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.01025390625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1152047475.0, "reward": 8.030361127853393, "reward_std": 0.30070753693580626, "rewards/combined_reward/mean": 8.03036093711853, "rewards/combined_reward/std": 2.2856724858283997, "step": 11250 }, { "clip_ratio/high_max": 0.002205286079455012, "clip_ratio/high_mean": 0.001287049333526132, "clip_ratio/low_mean": 0.0008338833232604278, "clip_ratio/low_min": 0.00026339896891537743, "clip_ratio/region_mean": 0.0021209326645475813, "completions/clipped_ratio": 0.0, "completions/max_length": 568.3, "completions/max_terminated_length": 568.3, "completions/mean_length": 208.2190139770508, "completions/mean_terminated_length": 208.2190139770508, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.03178556494725247, "epoch": 6.903581267217631, "frac_reward_zero_std": 0.1916666742414236, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1155147196.0, "reward": 8.156415700912476, "reward_std": 0.3948823392391205, "rewards/combined_reward/mean": 8.156415510177613, "rewards/combined_reward/std": 2.1471452713012695, "step": 11280 }, { "clip_ratio/high_max": 0.0020918658808416995, "clip_ratio/high_mean": 0.0012527590149450892, "clip_ratio/low_mean": 0.000700552420069774, "clip_ratio/low_min": 0.00016289934089096884, "clip_ratio/region_mean": 0.001953311423615863, "completions/clipped_ratio": 0.0, "completions/max_length": 561.7, "completions/max_terminated_length": 561.7, "completions/mean_length": 200.0835983276367, "completions/mean_terminated_length": 200.0835983276367, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.031630559964105485, "epoch": 6.921946740128559, "frac_reward_zero_std": 0.21666667312383653, "grad_norm": 0.026123046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1158215677.0, "reward": 7.95710391998291, "reward_std": 0.3564890742301941, "rewards/combined_reward/mean": 7.957103824615478, "rewards/combined_reward/std": 2.4821357250213625, "step": 11310 }, { "clip_ratio/high_max": 0.001652406987947567, "clip_ratio/high_mean": 0.0009813934877456632, "clip_ratio/low_mean": 0.0008384480518240404, "clip_ratio/low_min": 0.00027159954770468176, "clip_ratio/region_mean": 0.001819841550119842, "completions/clipped_ratio": 0.0, "completions/max_length": 595.7, "completions/max_terminated_length": 595.7, "completions/mean_length": 184.31068267822266, "completions/mean_terminated_length": 184.31068267822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028715881115446488, "epoch": 6.940312213039486, "frac_reward_zero_std": 0.15833333544433117, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1161223590.0, "reward": 7.758367919921875, "reward_std": 0.33900153189897536, "rewards/combined_reward/mean": 7.758367824554443, "rewards/combined_reward/std": 2.482869803905487, "step": 11340 }, { "clip_ratio/high_max": 0.0019877912981125217, "clip_ratio/high_mean": 0.0011451941980340052, "clip_ratio/low_mean": 0.0007213800173000588, "clip_ratio/low_min": 0.0002027524004612739, "clip_ratio/region_mean": 0.001866574221639894, "completions/clipped_ratio": 0.0, "completions/max_length": 566.2, "completions/max_terminated_length": 566.2, "completions/mean_length": 202.06745300292968, "completions/mean_terminated_length": 202.06745300292968, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02827616739086807, "epoch": 6.958677685950414, "frac_reward_zero_std": 0.1333333384245634, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1164299689.0, "reward": 8.075620698928834, "reward_std": 0.32244143187999724, "rewards/combined_reward/mean": 8.075620317459107, "rewards/combined_reward/std": 2.1139698386192323, "step": 11370 }, { "clip_ratio/high_max": 0.0019693530188912216, "clip_ratio/high_mean": 0.0011837147205369547, "clip_ratio/low_mean": 0.0008172493859698685, "clip_ratio/low_min": 0.000241065140774784, "clip_ratio/region_mean": 0.002000964104081504, "completions/clipped_ratio": 0.0, "completions/max_length": 530.4, "completions/max_terminated_length": 530.4, "completions/mean_length": 195.33255767822266, "completions/mean_terminated_length": 195.33255767822266, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.031030908226966858, "epoch": 6.97704315886134, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1167349926.0, "reward": 7.783879947662354, "reward_std": 0.3573672160506248, "rewards/combined_reward/mean": 7.783879661560059, "rewards/combined_reward/std": 2.590010237693787, "step": 11400 }, { "clip_ratio/high_max": 0.0020454163837712257, "clip_ratio/high_mean": 0.001149543735421806, "clip_ratio/low_mean": 0.0008164948261158619, "clip_ratio/low_min": 0.000207980791068015, "clip_ratio/region_mean": 0.0019660385592336145, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 209.05026550292968, "completions/mean_terminated_length": 209.05026550292968, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.030656565834457677, "epoch": 6.995408631772268, "frac_reward_zero_std": 0.15833333507180214, "grad_norm": 0.0267333984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1170452839.0, "reward": 7.957331228256225, "reward_std": 0.34142611026763914, "rewards/combined_reward/mean": 7.957331037521362, "rewards/combined_reward/std": 2.4075560212135314, "step": 11430 }, { "clip_ratio/high_max": 0.0026502252525848445, "clip_ratio/high_mean": 0.0014217249408829957, "clip_ratio/low_mean": 0.0009794809870907585, "clip_ratio/low_min": 0.00025666073860684563, "clip_ratio/region_mean": 0.002401205940080471, "completions/clipped_ratio": 0.0, "completions/max_length": 664.5, "completions/max_terminated_length": 664.5, "completions/mean_length": 202.47500610351562, "completions/mean_terminated_length": 202.47500610351562, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02879738338963419, "epoch": 7.014080195898377, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1173530503.0, "reward": 8.096567392349243, "reward_std": 0.33887441009283065, "rewards/combined_reward/mean": 8.096567153930664, "rewards/combined_reward/std": 2.0788731217384337, "step": 11460 }, { "clip_ratio/high_max": 0.0029238095667096786, "clip_ratio/high_mean": 0.0015764700692670885, "clip_ratio/low_mean": 0.001123921298130881, "clip_ratio/low_min": 0.0003238039108206673, "clip_ratio/region_mean": 0.00270039136618531, "completions/clipped_ratio": 0.0, "completions/max_length": 536.9, "completions/max_terminated_length": 536.9, "completions/mean_length": 204.41146545410157, "completions/mean_terminated_length": 204.41146545410157, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.03348758914507925, "epoch": 7.032445668809305, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1176615603.0, "reward": 7.868135643005371, "reward_std": 0.32045462280511855, "rewards/combined_reward/mean": 7.86813530921936, "rewards/combined_reward/std": 2.5668864965438845, "step": 11490 }, { "clip_ratio/high_max": 0.002705592167330906, "clip_ratio/high_mean": 0.001432499793372699, "clip_ratio/low_mean": 0.0009256957603308062, "clip_ratio/low_min": 0.00022483874636236578, "clip_ratio/region_mean": 0.002358195557947814, "completions/clipped_ratio": 0.0, "completions/max_length": 610.4, "completions/max_terminated_length": 610.4, "completions/mean_length": 209.58099365234375, "completions/mean_terminated_length": 209.58099365234375, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03002297176669041, "epoch": 7.050811141720232, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1179720554.0, "reward": 8.09481601715088, "reward_std": 0.37749641537666323, "rewards/combined_reward/mean": 8.094815635681153, "rewards/combined_reward/std": 2.295390176773071, "step": 11520 }, { "clip_ratio/high_max": 0.0023414910977104836, "clip_ratio/high_mean": 0.0013762663295589542, "clip_ratio/low_mean": 0.001047544572308349, "clip_ratio/low_min": 0.0003738601415534504, "clip_ratio/region_mean": 0.002423810905505282, "completions/clipped_ratio": 0.0, "completions/max_length": 610.9, "completions/max_terminated_length": 610.9, "completions/mean_length": 206.7877639770508, "completions/mean_terminated_length": 206.7877639770508, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.032301084309195476, "epoch": 7.06917661463116, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1182814779.0, "reward": 7.9934163093566895, "reward_std": 0.33926027566194533, "rewards/combined_reward/mean": 7.993416166305542, "rewards/combined_reward/std": 2.274701988697052, "step": 11550 }, { "clip_ratio/high_max": 0.0023109692481133, "clip_ratio/high_mean": 0.0013384988322892846, "clip_ratio/low_mean": 0.0009485708274951321, "clip_ratio/low_min": 0.00033238233275672733, "clip_ratio/region_mean": 0.002287069647475922, "completions/clipped_ratio": 0.0, "completions/max_length": 550.3, "completions/max_terminated_length": 550.3, "completions/mean_length": 202.50625457763672, "completions/mean_terminated_length": 202.50625457763672, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029644800916624567, "epoch": 7.087542087542087, "frac_reward_zero_std": 0.20000000447034835, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1185892563.0, "reward": 8.103674030303955, "reward_std": 0.29299232363700867, "rewards/combined_reward/mean": 8.103673887252807, "rewards/combined_reward/std": 2.335701072216034, "step": 11580 }, { "clip_ratio/high_max": 0.0020830881912843324, "clip_ratio/high_mean": 0.0011459807621577057, "clip_ratio/low_mean": 0.0008634760285834394, "clip_ratio/low_min": 0.0002906728766295904, "clip_ratio/region_mean": 0.002009456796319379, "completions/clipped_ratio": 0.0, "completions/max_length": 573.2, "completions/max_terminated_length": 573.2, "completions/mean_length": 197.66875762939452, "completions/mean_terminated_length": 197.66875762939452, "completions/min_length": 45.5, "completions/min_terminated_length": 45.5, "entropy": 0.02513511963188648, "epoch": 7.105907560453015, "frac_reward_zero_std": 0.17083333991467953, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1188951771.0, "reward": 8.025763845443725, "reward_std": 0.3750196024775505, "rewards/combined_reward/mean": 8.025763845443725, "rewards/combined_reward/std": 2.2885470271110533, "step": 11610 }, { "clip_ratio/high_max": 0.0020161242469233305, "clip_ratio/high_mean": 0.0011255765610258095, "clip_ratio/low_mean": 0.0009349598329815005, "clip_ratio/low_min": 0.0002477395287617886, "clip_ratio/region_mean": 0.002060536414986321, "completions/clipped_ratio": 0.0, "completions/max_length": 578.2, "completions/max_terminated_length": 578.2, "completions/mean_length": 199.1994842529297, "completions/mean_terminated_length": 199.1994842529297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02589169074781239, "epoch": 7.124273033363942, "frac_reward_zero_std": 0.1833333417773247, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1192016857.0, "reward": 8.263413763046264, "reward_std": 0.2998686835169792, "rewards/combined_reward/mean": 8.263413619995116, "rewards/combined_reward/std": 2.083854651451111, "step": 11640 }, { "clip_ratio/high_max": 0.002379444329077766, "clip_ratio/high_mean": 0.001400882662346703, "clip_ratio/low_mean": 0.0009248764847143321, "clip_ratio/low_min": 0.00023615026536087194, "clip_ratio/region_mean": 0.002325759140755205, "completions/clipped_ratio": 0.0, "completions/max_length": 683.4, "completions/max_terminated_length": 683.4, "completions/mean_length": 204.1406280517578, "completions/mean_terminated_length": 204.1406280517578, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.032075514628862344, "epoch": 7.14263850627487, "frac_reward_zero_std": 0.17500000819563866, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1195100917.0, "reward": 8.164218425750732, "reward_std": 0.3821890473365784, "rewards/combined_reward/mean": 8.164218139648437, "rewards/combined_reward/std": 2.1902728199958803, "step": 11670 }, { "clip_ratio/high_max": 0.0029378978729558487, "clip_ratio/high_mean": 0.0017523231021186803, "clip_ratio/low_mean": 0.0010764756514011727, "clip_ratio/low_min": 0.0003063696962878263, "clip_ratio/region_mean": 0.002828798743818576, "completions/clipped_ratio": 0.0, "completions/max_length": 666.3, "completions/max_terminated_length": 666.3, "completions/mean_length": 201.81250457763673, "completions/mean_terminated_length": 201.81250457763673, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03575697573833168, "epoch": 7.161003979185797, "frac_reward_zero_std": 0.20000000558793546, "grad_norm": 0.01434326171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1198176037.0, "reward": 8.025641059875488, "reward_std": 0.353277350962162, "rewards/combined_reward/mean": 8.025640821456909, "rewards/combined_reward/std": 2.1672765970230103, "step": 11700 }, { "clip_ratio/high_max": 0.002696199458053646, "clip_ratio/high_mean": 0.0015031025150771407, "clip_ratio/low_mean": 0.0010392264140439997, "clip_ratio/low_min": 0.000297625260282075, "clip_ratio/region_mean": 0.00254232893542697, "completions/clipped_ratio": 0.0, "completions/max_length": 563.9, "completions/max_terminated_length": 563.9, "completions/mean_length": 187.62161865234376, "completions/mean_terminated_length": 187.62161865234376, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03223749155489107, "epoch": 7.179369452096725, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1201196664.0, "reward": 7.611458158493042, "reward_std": 0.3205262079834938, "rewards/combined_reward/mean": 7.611457681655883, "rewards/combined_reward/std": 2.796560215950012, "step": 11730 }, { "clip_ratio/high_max": 0.0023753060013405047, "clip_ratio/high_mean": 0.001342077243558985, "clip_ratio/low_mean": 0.0008663514783014155, "clip_ratio/low_min": 0.00023858298518462107, "clip_ratio/region_mean": 0.0022084287144631767, "completions/clipped_ratio": 0.0, "completions/max_length": 612.5, "completions/max_terminated_length": 612.5, "completions/mean_length": 199.3765670776367, "completions/mean_terminated_length": 199.3765670776367, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.02643754123710096, "epoch": 7.197734925007652, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1204262430.0, "reward": 8.026914405822755, "reward_std": 0.3339911252260208, "rewards/combined_reward/mean": 8.02691421508789, "rewards/combined_reward/std": 2.4534616470336914, "step": 11760 }, { "clip_ratio/high_max": 0.0026917531717723857, "clip_ratio/high_mean": 0.0015270276591763832, "clip_ratio/low_mean": 0.001015149470428393, "clip_ratio/low_min": 0.0002842429118269744, "clip_ratio/region_mean": 0.002542177129847308, "completions/clipped_ratio": 0.0, "completions/max_length": 523.6, "completions/max_terminated_length": 523.6, "completions/mean_length": 200.70209045410155, "completions/mean_terminated_length": 200.70209045410155, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.030355841387063265, "epoch": 7.21610039791858, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0089111328125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1207333286.0, "reward": 8.132734251022338, "reward_std": 0.2830321803689003, "rewards/combined_reward/mean": 8.132733964920044, "rewards/combined_reward/std": 2.169464576244354, "step": 11790 }, { "clip_ratio/high_max": 0.0023617433456820437, "clip_ratio/high_mean": 0.0013430797514350464, "clip_ratio/low_mean": 0.0009823968467874997, "clip_ratio/low_min": 0.0002267974407004658, "clip_ratio/region_mean": 0.0023254766080450886, "completions/clipped_ratio": 0.0, "completions/max_length": 551.8, "completions/max_terminated_length": 551.8, "completions/mean_length": 198.83281402587892, "completions/mean_terminated_length": 198.83281402587892, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.029841861004630724, "epoch": 7.234465870829507, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.005340576171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1210396964.0, "reward": 7.989105129241944, "reward_std": 0.2835030764341354, "rewards/combined_reward/mean": 7.989104700088501, "rewards/combined_reward/std": 2.307388162612915, "step": 11820 }, { "clip_ratio/high_max": 0.0023181833079434, "clip_ratio/high_mean": 0.0013388077063912836, "clip_ratio/low_mean": 0.001085155666078208, "clip_ratio/low_min": 0.00038881567622108074, "clip_ratio/region_mean": 0.0024239633746522787, "completions/clipped_ratio": 0.0, "completions/max_length": 632.9, "completions/max_terminated_length": 632.9, "completions/mean_length": 198.80859985351563, "completions/mean_terminated_length": 198.80859985351563, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.029715709869439402, "epoch": 7.252831343740435, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 1213460549.0, "reward": 8.178104209899903, "reward_std": 0.30901400819420816, "rewards/combined_reward/mean": 8.178104305267334, "rewards/combined_reward/std": 2.1466352462768556, "step": 11850 }, { "clip_ratio/high_max": 0.0022903848133864813, "clip_ratio/high_mean": 0.0013517204558108158, "clip_ratio/low_mean": 0.0008822225529002026, "clip_ratio/low_min": 0.0001929854421177879, "clip_ratio/region_mean": 0.00223394300264772, "completions/clipped_ratio": 0.0, "completions/max_length": 583.9, "completions/max_terminated_length": 583.9, "completions/mean_length": 204.64505615234376, "completions/mean_terminated_length": 204.64505615234376, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.029350959059471885, "epoch": 7.271196816651362, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1216546546.0, "reward": 8.289608001708984, "reward_std": 0.3270338147878647, "rewards/combined_reward/mean": 8.289607429504395, "rewards/combined_reward/std": 1.9634893655776977, "step": 11880 }, { "clip_ratio/high_max": 0.002671030330869447, "clip_ratio/high_mean": 0.0013957804194433265, "clip_ratio/low_mean": 0.0009767464329343057, "clip_ratio/low_min": 0.00020429251865910675, "clip_ratio/region_mean": 0.0023725268702643612, "completions/clipped_ratio": 0.0, "completions/max_length": 492.2, "completions/max_terminated_length": 492.2, "completions/mean_length": 202.62734832763672, "completions/mean_terminated_length": 202.62734832763672, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029946652598058184, "epoch": 7.28956228956229, "frac_reward_zero_std": 0.1250000037252903, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0027, "num_tokens": 1219624795.0, "reward": 8.042465591430664, "reward_std": 0.42946452349424363, "rewards/combined_reward/mean": 8.042465114593506, "rewards/combined_reward/std": 2.254536533355713, "step": 11910 }, { "clip_ratio/high_max": 0.0022931075383288164, "clip_ratio/high_mean": 0.0012722229006612907, "clip_ratio/low_mean": 0.0009725304818857694, "clip_ratio/low_min": 0.00029055862372236635, "clip_ratio/region_mean": 0.0022447533886103582, "completions/clipped_ratio": 0.0, "completions/max_length": 644.2, "completions/max_terminated_length": 644.2, "completions/mean_length": 189.77448425292968, "completions/mean_terminated_length": 189.77448425292968, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.026385789271444082, "epoch": 7.307927762473217, "frac_reward_zero_std": 0.2041666705161333, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1222653689.0, "reward": 8.23063645362854, "reward_std": 0.32033751606941224, "rewards/combined_reward/mean": 8.230636405944825, "rewards/combined_reward/std": 2.0065919756889343, "step": 11940 }, { "clip_ratio/high_max": 0.0023723165393069695, "clip_ratio/high_mean": 0.0012915080641202317, "clip_ratio/low_mean": 0.0008664956296343008, "clip_ratio/low_min": 0.00023821324696958376, "clip_ratio/region_mean": 0.0021580037071544213, "completions/clipped_ratio": 0.0, "completions/max_length": 578.7, "completions/max_terminated_length": 578.7, "completions/mean_length": 203.53802490234375, "completions/mean_terminated_length": 203.53802490234375, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.030692892940714957, "epoch": 7.326293235384145, "frac_reward_zero_std": 0.12916666977107524, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1225735435.0, "reward": 7.865723419189453, "reward_std": 0.35255300998687744, "rewards/combined_reward/mean": 7.865723085403443, "rewards/combined_reward/std": 2.5746230363845823, "step": 11970 }, { "clip_ratio/high_max": 0.0025354325960506687, "clip_ratio/high_mean": 0.0014290896038194963, "clip_ratio/low_mean": 0.0010306022374910147, "clip_ratio/low_min": 0.00033267943387424264, "clip_ratio/region_mean": 0.0024596918451910217, "completions/clipped_ratio": 0.0, "completions/max_length": 497.7, "completions/max_terminated_length": 497.7, "completions/mean_length": 200.4987030029297, "completions/mean_terminated_length": 200.4987030029297, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.028904926683753728, "epoch": 7.344658708295072, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1228805510.0, "reward": 8.071453189849853, "reward_std": 0.24950546771287918, "rewards/combined_reward/mean": 8.071452951431274, "rewards/combined_reward/std": 2.3576292753219605, "step": 12000 }, { "clip_ratio/high_max": 0.002484371404959044, "clip_ratio/high_mean": 0.0014222705267457059, "clip_ratio/low_mean": 0.0009619890988687985, "clip_ratio/low_min": 0.00021838904067408293, "clip_ratio/region_mean": 0.002384259641985409, "completions/clipped_ratio": 0.0, "completions/max_length": 619.5, "completions/max_terminated_length": 619.5, "completions/mean_length": 202.85521240234374, "completions/mean_terminated_length": 202.85521240234374, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.029808748839423062, "epoch": 7.363024181206, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1231884634.0, "reward": 8.133028888702393, "reward_std": 0.36201013922691344, "rewards/combined_reward/mean": 8.133028268814087, "rewards/combined_reward/std": 2.143662416934967, "step": 12030 }, { "clip_ratio/high_max": 0.002284809144718262, "clip_ratio/high_mean": 0.0012434593359406183, "clip_ratio/low_mean": 0.0009359046125609893, "clip_ratio/low_min": 0.0002316986079677008, "clip_ratio/region_mean": 0.0021793639607494697, "completions/clipped_ratio": 0.0, "completions/max_length": 528.1, "completions/max_terminated_length": 528.1, "completions/mean_length": 205.8156295776367, "completions/mean_terminated_length": 205.8156295776367, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.029709018021821975, "epoch": 7.381389654116926, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0096435546875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1234975126.0, "reward": 8.228102350234986, "reward_std": 0.31543356776237486, "rewards/combined_reward/mean": 8.228102207183838, "rewards/combined_reward/std": 2.005012500286102, "step": 12060 }, { "clip_ratio/high_max": 0.002796630342587984, "clip_ratio/high_mean": 0.0014604971162649842, "clip_ratio/low_mean": 0.0010565641254894824, "clip_ratio/low_min": 0.0002958134265403108, "clip_ratio/region_mean": 0.0025170612296885036, "completions/clipped_ratio": 0.0, "completions/max_length": 669.3, "completions/max_terminated_length": 669.3, "completions/mean_length": 198.2750045776367, "completions/mean_terminated_length": 198.2750045776367, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.028683760420729718, "epoch": 7.399755127027854, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.00592041015625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1238036662.0, "reward": 7.918029165267944, "reward_std": 0.38823519051074984, "rewards/combined_reward/mean": 7.918029069900513, "rewards/combined_reward/std": 2.337974154949188, "step": 12090 }, { "clip_ratio/high_max": 0.0025613632814687056, "clip_ratio/high_mean": 0.0014330986815669652, "clip_ratio/low_mean": 0.0009816474655356918, "clip_ratio/low_min": 0.0003535349026302962, "clip_ratio/region_mean": 0.0024147461333389702, "completions/clipped_ratio": 0.0, "completions/max_length": 659.1, "completions/max_terminated_length": 659.1, "completions/mean_length": 204.31302490234376, "completions/mean_terminated_length": 204.31302490234376, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03029924977260331, "epoch": 7.418120599938781, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1241121384.0, "reward": 7.961383438110351, "reward_std": 0.35526159554719927, "rewards/combined_reward/mean": 7.961383104324341, "rewards/combined_reward/std": 2.377382278442383, "step": 12120 }, { "clip_ratio/high_max": 0.002162910907403178, "clip_ratio/high_mean": 0.0011840208603340822, "clip_ratio/low_mean": 0.0008004958907501229, "clip_ratio/low_min": 0.00019130979829545442, "clip_ratio/region_mean": 0.001984516744172045, "completions/clipped_ratio": 0.0, "completions/max_length": 557.3, "completions/max_terminated_length": 557.3, "completions/mean_length": 192.8807342529297, "completions/mean_terminated_length": 192.8807342529297, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.025572876213118434, "epoch": 7.436486072849709, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1244162206.0, "reward": 8.208733510971069, "reward_std": 0.36583987027406695, "rewards/combined_reward/mean": 8.20873303413391, "rewards/combined_reward/std": 2.194824731349945, "step": 12150 }, { "clip_ratio/high_max": 0.0022326137472797806, "clip_ratio/high_mean": 0.0012885927933287652, "clip_ratio/low_mean": 0.0009252400285428545, "clip_ratio/low_min": 0.00021985245451408748, "clip_ratio/region_mean": 0.0022138328262371943, "completions/clipped_ratio": 0.0, "completions/max_length": 609.6, "completions/max_terminated_length": 609.6, "completions/mean_length": 202.9471420288086, "completions/mean_terminated_length": 202.9471420288086, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02801849174623688, "epoch": 7.454851545760636, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.04052734375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1247241683.0, "reward": 8.077804374694825, "reward_std": 0.37410866022109984, "rewards/combined_reward/mean": 8.077804374694825, "rewards/combined_reward/std": 2.395075595378876, "step": 12180 }, { "clip_ratio/high_max": 0.0027772094144893344, "clip_ratio/high_mean": 0.0015459577512956456, "clip_ratio/low_mean": 0.000873410851030106, "clip_ratio/low_min": 0.00020349880166274184, "clip_ratio/region_mean": 0.0024193685979601773, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 198.55912170410156, "completions/mean_terminated_length": 198.55912170410156, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02811054027018448, "epoch": 7.473217018671564, "frac_reward_zero_std": 0.20000000298023224, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1250304310.0, "reward": 7.922275876998901, "reward_std": 0.3236138761043549, "rewards/combined_reward/mean": 7.92227578163147, "rewards/combined_reward/std": 2.431255042552948, "step": 12210 }, { "clip_ratio/high_max": 0.002124889531114604, "clip_ratio/high_mean": 0.0011893674076418393, "clip_ratio/low_mean": 0.0010170395032522114, "clip_ratio/low_min": 0.0003951857700788726, "clip_ratio/region_mean": 0.0022064069077411355, "completions/clipped_ratio": 0.0, "completions/max_length": 578.5, "completions/max_terminated_length": 578.5, "completions/mean_length": 202.0171920776367, "completions/mean_terminated_length": 202.0171920776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.026870010195610423, "epoch": 7.491582491582491, "frac_reward_zero_std": 0.20416667200624944, "grad_norm": 0.024658203125, "learning_rate": 5e-06, "loss": 0.0029, "num_tokens": 1253380216.0, "reward": 8.02344765663147, "reward_std": 0.39275089651346207, "rewards/combined_reward/mean": 8.02344732284546, "rewards/combined_reward/std": 2.2780104637145997, "step": 12240 }, { "clip_ratio/high_max": 0.002261257639232402, "clip_ratio/high_mean": 0.0013177741042454728, "clip_ratio/low_mean": 0.0010110117293758473, "clip_ratio/low_min": 0.0002936079690698534, "clip_ratio/region_mean": 0.002328785836289171, "completions/clipped_ratio": 0.0, "completions/max_length": 497.1, "completions/max_terminated_length": 497.1, "completions/mean_length": 200.0846405029297, "completions/mean_terminated_length": 200.0846405029297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026901623886078595, "epoch": 7.509947964493419, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1256448701.0, "reward": 8.060944271087646, "reward_std": 0.3624969214200974, "rewards/combined_reward/mean": 8.060944271087646, "rewards/combined_reward/std": 2.4682740449905394, "step": 12270 }, { "clip_ratio/high_max": 0.002188401466507154, "clip_ratio/high_mean": 0.0012915287358434095, "clip_ratio/low_mean": 0.0009266928719929031, "clip_ratio/low_min": 0.00031252710323315114, "clip_ratio/region_mean": 0.002218221616446196, "completions/clipped_ratio": 0.0, "completions/max_length": 567.1, "completions/max_terminated_length": 567.1, "completions/mean_length": 207.1440170288086, "completions/mean_terminated_length": 207.1440170288086, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.029056017519906163, "epoch": 7.528313437404346, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.0123291015625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1259544294.0, "reward": 8.054662084579467, "reward_std": 0.36447775214910505, "rewards/combined_reward/mean": 8.05466160774231, "rewards/combined_reward/std": 2.1594608426094055, "step": 12300 }, { "clip_ratio/high_max": 0.0030053654743824156, "clip_ratio/high_mean": 0.0017623394979940107, "clip_ratio/low_mean": 0.0010968124720723911, "clip_ratio/low_min": 0.00027935645266552456, "clip_ratio/region_mean": 0.0028591520046272003, "completions/clipped_ratio": 0.0, "completions/max_length": 602.6, "completions/max_terminated_length": 602.6, "completions/mean_length": 199.69766082763672, "completions/mean_terminated_length": 199.69766082763672, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.033780418786530694, "epoch": 7.546678910315274, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1262611293.0, "reward": 8.046434593200683, "reward_std": 0.34798331558704376, "rewards/combined_reward/mean": 8.046434307098389, "rewards/combined_reward/std": 2.333920049667358, "step": 12330 }, { "clip_ratio/high_max": 0.002573879273647132, "clip_ratio/high_mean": 0.0014966601615014953, "clip_ratio/low_mean": 0.0011723071222756213, "clip_ratio/low_min": 0.0003493445049874329, "clip_ratio/region_mean": 0.0026689672978439678, "completions/clipped_ratio": 0.0, "completions/max_length": 512.4, "completions/max_terminated_length": 512.4, "completions/mean_length": 195.2968780517578, "completions/mean_terminated_length": 195.2968780517578, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03112768760571877, "epoch": 7.565044383226201, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1265661393.0, "reward": 8.02637677192688, "reward_std": 0.31073973178863523, "rewards/combined_reward/mean": 8.026376724243164, "rewards/combined_reward/std": 2.342959129810333, "step": 12360 }, { "clip_ratio/high_max": 0.0021965974718720343, "clip_ratio/high_mean": 0.0012964099708672923, "clip_ratio/low_mean": 0.0008542761377611896, "clip_ratio/low_min": 0.00020243001223813433, "clip_ratio/region_mean": 0.00215068610208012, "completions/clipped_ratio": 0.0, "completions/max_length": 550.6, "completions/max_terminated_length": 550.6, "completions/mean_length": 185.18203430175782, "completions/mean_terminated_length": 185.18203430175782, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.026959034567698836, "epoch": 7.583409856137129, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1268672652.0, "reward": 7.881915760040283, "reward_std": 0.2966366454958916, "rewards/combined_reward/mean": 7.881915760040283, "rewards/combined_reward/std": 2.5801676630973818, "step": 12390 }, { "clip_ratio/high_max": 0.0017666247335000663, "clip_ratio/high_mean": 0.001007680307884584, "clip_ratio/low_mean": 0.0008325957591296173, "clip_ratio/low_min": 0.00016524608436156995, "clip_ratio/region_mean": 0.001840276059374446, "completions/clipped_ratio": 0.0, "completions/max_length": 583.8571428571429, "completions/max_terminated_length": 583.8571428571429, "completions/mean_length": 213.57440839494978, "completions/mean_terminated_length": 213.57440839494978, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.030020014801993965, "epoch": 7.601163146617692, "frac_reward_zero_std": 0.12500000425747462, "grad_norm": 0.0155029296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1271783359.0, "reward": 8.200811862945557, "reward_std": 0.33478125078337534, "rewards/combined_reward/mean": 8.200811454227992, "rewards/combined_reward/std": 2.1399766887937273, "step": 12420 }, { "clip_ratio/high_max": 0.0023460850682264817, "clip_ratio/high_mean": 0.0013067240882567905, "clip_ratio/low_mean": 0.0008702122909501971, "clip_ratio/low_min": 0.0002775719711886874, "clip_ratio/region_mean": 0.0021769363796920516, "completions/clipped_ratio": 0.0, "completions/max_length": 562.5, "completions/max_terminated_length": 562.5, "completions/mean_length": 189.5510467529297, "completions/mean_terminated_length": 189.5510467529297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03191234767436981, "epoch": 7.619528619528619, "frac_reward_zero_std": 0.2458333417773247, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1274811395.0, "reward": 7.801935529708862, "reward_std": 0.29824732691049577, "rewards/combined_reward/mean": 7.801935386657715, "rewards/combined_reward/std": 2.489002287387848, "step": 12450 }, { "clip_ratio/high_max": 0.001965081759165817, "clip_ratio/high_mean": 0.0011262542997428683, "clip_ratio/low_mean": 0.0008232178549709109, "clip_ratio/low_min": 0.0002878861574572511, "clip_ratio/region_mean": 0.0019494721554413749, "completions/clipped_ratio": 0.0, "completions/max_length": 518.3, "completions/max_terminated_length": 518.3, "completions/mean_length": 203.89453735351563, "completions/mean_terminated_length": 203.89453735351563, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.028290740695471564, "epoch": 7.637894092439547, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.01434326171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1277894510.0, "reward": 7.986710500717163, "reward_std": 0.3230353891849518, "rewards/combined_reward/mean": 7.986710500717163, "rewards/combined_reward/std": 2.232128918170929, "step": 12480 }, { "clip_ratio/high_max": 0.0019388386465531463, "clip_ratio/high_mean": 0.0010885462797887158, "clip_ratio/low_mean": 0.0008225991363966993, "clip_ratio/low_min": 0.00024685660367443536, "clip_ratio/region_mean": 0.0019111454404386072, "completions/clipped_ratio": 0.0, "completions/max_length": 586.8, "completions/max_terminated_length": 586.8, "completions/mean_length": 196.59922485351564, "completions/mean_terminated_length": 196.59922485351564, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029408722603693604, "epoch": 7.656259565350474, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.0064697265625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1280949611.0, "reward": 8.185135793685912, "reward_std": 0.28906214982271194, "rewards/combined_reward/mean": 8.185135221481323, "rewards/combined_reward/std": 2.2751994252204897, "step": 12510 }, { "clip_ratio/high_max": 0.0018360523439090077, "clip_ratio/high_mean": 0.0010461318312688188, "clip_ratio/low_mean": 0.0008074978672084398, "clip_ratio/low_min": 0.0002382904337234019, "clip_ratio/region_mean": 0.0018536296973858649, "completions/clipped_ratio": 0.0, "completions/max_length": 552.4, "completions/max_terminated_length": 552.4, "completions/mean_length": 188.71250610351564, "completions/mean_terminated_length": 188.71250610351564, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.02839336011869212, "epoch": 7.674625038261402, "frac_reward_zero_std": 0.21250000596046448, "grad_norm": 0.029052734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1283974427.0, "reward": 7.884507942199707, "reward_std": 0.3498475641012192, "rewards/combined_reward/mean": 7.884507656097412, "rewards/combined_reward/std": 2.3734697580337523, "step": 12540 }, { "clip_ratio/high_max": 0.0018479983884996424, "clip_ratio/high_mean": 0.0010295049777293266, "clip_ratio/low_mean": 0.0008056957679097347, "clip_ratio/low_min": 0.00023198957545294736, "clip_ratio/region_mean": 0.001835200731390311, "completions/clipped_ratio": 0.0, "completions/max_length": 527.1, "completions/max_terminated_length": 527.1, "completions/mean_length": 181.45990142822265, "completions/mean_terminated_length": 181.45990142822265, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.029785606916993855, "epoch": 7.692990511172329, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1286971393.0, "reward": 8.125857019424439, "reward_std": 0.2921998590230942, "rewards/combined_reward/mean": 8.125856828689574, "rewards/combined_reward/std": 2.2074210822582243, "step": 12570 }, { "clip_ratio/high_max": 0.002293233194601877, "clip_ratio/high_mean": 0.0013446473322498301, "clip_ratio/low_mean": 0.000764712160162162, "clip_ratio/low_min": 0.00018522441775227586, "clip_ratio/region_mean": 0.002109359494837311, "completions/clipped_ratio": 0.0, "completions/max_length": 666.2, "completions/max_terminated_length": 666.2, "completions/mean_length": 207.45000762939452, "completions/mean_terminated_length": 207.45000762939452, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.032732095460717876, "epoch": 7.711355984083257, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.01116943359375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1290068161.0, "reward": 7.911301898956299, "reward_std": 0.3728142619132996, "rewards/combined_reward/mean": 7.911301755905152, "rewards/combined_reward/std": 2.543099117279053, "step": 12600 }, { "clip_ratio/high_max": 0.0016291995818998354, "clip_ratio/high_mean": 0.0009749018867296399, "clip_ratio/low_mean": 0.0007171475185411206, "clip_ratio/low_min": 0.00022792629897594452, "clip_ratio/region_mean": 0.0016920494090300053, "completions/clipped_ratio": 0.0, "completions/max_length": 597.1, "completions/max_terminated_length": 597.1, "completions/mean_length": 205.3375045776367, "completions/mean_terminated_length": 205.3375045776367, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.02737361566784481, "epoch": 7.729721456994184, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1293156817.0, "reward": 8.004576015472413, "reward_std": 0.30282439216971396, "rewards/combined_reward/mean": 8.004575634002686, "rewards/combined_reward/std": 2.2950407385826113, "step": 12630 }, { "clip_ratio/high_max": 0.0016375704974052497, "clip_ratio/high_mean": 0.0009963634103769437, "clip_ratio/low_mean": 0.0007811291532561881, "clip_ratio/low_min": 0.00024094398443897565, "clip_ratio/region_mean": 0.0017774925564784402, "completions/clipped_ratio": 0.0, "completions/max_length": 516.6, "completions/max_terminated_length": 516.6, "completions/mean_length": 200.3557342529297, "completions/mean_terminated_length": 200.3557342529297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02727320611787339, "epoch": 7.748086929905112, "frac_reward_zero_std": 0.2125000063329935, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1296226343.0, "reward": 8.140404319763183, "reward_std": 0.3101580113172531, "rewards/combined_reward/mean": 8.14040403366089, "rewards/combined_reward/std": 2.0987732291221617, "step": 12660 }, { "clip_ratio/high_max": 0.0020505829740917153, "clip_ratio/high_mean": 0.001085252075184447, "clip_ratio/low_mean": 0.0007363745899056084, "clip_ratio/low_min": 0.00017132940362595643, "clip_ratio/region_mean": 0.0018216266655751193, "completions/clipped_ratio": 0.0, "completions/max_length": 606.8, "completions/max_terminated_length": 606.8, "completions/mean_length": 189.40859680175782, "completions/mean_terminated_length": 189.40859680175782, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03383834493967394, "epoch": 7.766452402816039, "frac_reward_zero_std": 0.20416666902601718, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1299253832.0, "reward": 8.222685050964355, "reward_std": 0.2588932834565639, "rewards/combined_reward/mean": 8.222684478759765, "rewards/combined_reward/std": 2.092503106594086, "step": 12690 }, { "clip_ratio/high_max": 0.0018125865021526503, "clip_ratio/high_mean": 0.001051998450323784, "clip_ratio/low_mean": 0.0007742251217374966, "clip_ratio/low_min": 0.00019224332985080158, "clip_ratio/region_mean": 0.0018262235772757171, "completions/clipped_ratio": 0.0, "completions/max_length": 577.5, "completions/max_terminated_length": 577.5, "completions/mean_length": 194.8268295288086, "completions/mean_terminated_length": 194.8268295288086, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.02869369660814603, "epoch": 7.784817875726967, "frac_reward_zero_std": 0.14166666939854622, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1302302127.0, "reward": 8.3833083152771, "reward_std": 0.32079113870859144, "rewards/combined_reward/mean": 8.38330807685852, "rewards/combined_reward/std": 2.0240936636924745, "step": 12720 }, { "clip_ratio/high_max": 0.001900095345141987, "clip_ratio/high_mean": 0.0010600482276155768, "clip_ratio/low_mean": 0.0006289975178636572, "clip_ratio/low_min": 0.00016025705602563297, "clip_ratio/region_mean": 0.0016890457574239312, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 207.04687957763673, "completions/mean_terminated_length": 207.04687957763673, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.02758165275057157, "epoch": 7.803183348637894, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 1305397347.0, "reward": 8.13090901374817, "reward_std": 0.3474277496337891, "rewards/combined_reward/mean": 8.130908966064453, "rewards/combined_reward/std": 2.3017353892326353, "step": 12750 }, { "clip_ratio/high_max": 0.001995024896071603, "clip_ratio/high_mean": 0.0012261382704309653, "clip_ratio/low_mean": 0.0008233804067761715, "clip_ratio/low_min": 0.00017166627658298238, "clip_ratio/region_mean": 0.0020495186933355094, "completions/clipped_ratio": 0.0, "completions/max_length": 620.9, "completions/max_terminated_length": 620.9, "completions/mean_length": 200.8268280029297, "completions/mean_terminated_length": 200.8268280029297, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.029927697964012623, "epoch": 7.821548821548822, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 1308468682.0, "reward": 8.106758260726929, "reward_std": 0.381988936662674, "rewards/combined_reward/mean": 8.106757974624633, "rewards/combined_reward/std": 2.1744617104530333, "step": 12780 }, { "clip_ratio/high_max": 0.001923952955015314, "clip_ratio/high_mean": 0.0011439723870959521, "clip_ratio/low_mean": 0.0008357924739054094, "clip_ratio/low_min": 0.0002442560828058049, "clip_ratio/region_mean": 0.0019797648492385633, "completions/clipped_ratio": 0.0, "completions/max_length": 642.5, "completions/max_terminated_length": 642.5, "completions/mean_length": 195.5229232788086, "completions/mean_terminated_length": 195.5229232788086, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.030246459149445095, "epoch": 7.839914294459749, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1311519650.0, "reward": 7.975250625610352, "reward_std": 0.3649938553571701, "rewards/combined_reward/mean": 7.9752504348754885, "rewards/combined_reward/std": 2.263340675830841, "step": 12810 }, { "clip_ratio/high_max": 0.0018057453470343412, "clip_ratio/high_mean": 0.0011041986486816312, "clip_ratio/low_mean": 0.0007135603174295587, "clip_ratio/low_min": 0.00016287175239995121, "clip_ratio/region_mean": 0.0018177589634433388, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 194.8346405029297, "completions/mean_terminated_length": 194.8346405029297, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.028414623284091552, "epoch": 7.858279767370677, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.006500244140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1314567975.0, "reward": 8.216027641296387, "reward_std": 0.31808760315179824, "rewards/combined_reward/mean": 8.216027069091798, "rewards/combined_reward/std": 2.12193900346756, "step": 12840 }, { "clip_ratio/high_max": 0.0019958012611217177, "clip_ratio/high_mean": 0.001116109734478717, "clip_ratio/low_mean": 0.0007892675093292685, "clip_ratio/low_min": 0.00019802928569940075, "clip_ratio/region_mean": 0.001905377221798214, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 204.79349517822266, "completions/mean_terminated_length": 204.79349517822266, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03391778930090368, "epoch": 7.876645240281604, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.004852294921875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1317654542.0, "reward": 8.287318325042724, "reward_std": 0.34214196130633356, "rewards/combined_reward/mean": 8.28731803894043, "rewards/combined_reward/std": 2.0230463981628417, "step": 12870 }, { "clip_ratio/high_max": 0.0023856286560961356, "clip_ratio/high_mean": 0.0013832746037223842, "clip_ratio/low_mean": 0.0007756992376622899, "clip_ratio/low_min": 0.00028624545520870013, "clip_ratio/region_mean": 0.0021589738617573554, "completions/clipped_ratio": 0.0, "completions/max_length": 551.6, "completions/max_terminated_length": 551.6, "completions/mean_length": 193.7791732788086, "completions/mean_terminated_length": 193.7791732788086, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03139976772169272, "epoch": 7.895010713192532, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1320698814.0, "reward": 7.918189668655396, "reward_std": 0.3929837003350258, "rewards/combined_reward/mean": 7.918189477920532, "rewards/combined_reward/std": 2.3530362844467163, "step": 12900 }, { "clip_ratio/high_max": 0.0017950506570438543, "clip_ratio/high_mean": 0.0011065369269393462, "clip_ratio/low_mean": 0.0007506605060674094, "clip_ratio/low_min": 0.00022500803800843034, "clip_ratio/region_mean": 0.0018571974214864895, "completions/clipped_ratio": 0.0, "completions/max_length": 608.5, "completions/max_terminated_length": 608.5, "completions/mean_length": 208.3151077270508, "completions/mean_terminated_length": 208.3151077270508, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03131982437334955, "epoch": 7.913376186103459, "frac_reward_zero_std": 0.18750000409781933, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1323798904.0, "reward": 8.166026306152343, "reward_std": 0.3782848432660103, "rewards/combined_reward/mean": 8.166026067733764, "rewards/combined_reward/std": 2.1711834192276003, "step": 12930 }, { "clip_ratio/high_max": 0.0020733777254160184, "clip_ratio/high_mean": 0.0012921734902192838, "clip_ratio/low_mean": 0.0008018690640407537, "clip_ratio/low_min": 0.0002822282483975869, "clip_ratio/region_mean": 0.002094042544194963, "completions/clipped_ratio": 0.0, "completions/max_length": 524.3, "completions/max_terminated_length": 524.3, "completions/mean_length": 188.4869842529297, "completions/mean_terminated_length": 188.4869842529297, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.030229988787323237, "epoch": 7.931741659014387, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.006927490234375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1326822854.0, "reward": 7.860158157348633, "reward_std": 0.3177006497979164, "rewards/combined_reward/mean": 7.860157918930054, "rewards/combined_reward/std": 2.522070074081421, "step": 12960 }, { "clip_ratio/high_max": 0.0017672854871004044, "clip_ratio/high_mean": 0.0010287335697891345, "clip_ratio/low_mean": 0.0008378624300045582, "clip_ratio/low_min": 0.0002373666286681934, "clip_ratio/region_mean": 0.0018665960276848636, "completions/clipped_ratio": 0.0, "completions/max_length": 604.3, "completions/max_terminated_length": 604.3, "completions/mean_length": 195.35052490234375, "completions/mean_terminated_length": 195.35052490234375, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029319660536323986, "epoch": 7.950107131925313, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1329873160.0, "reward": 7.862717771530152, "reward_std": 0.34810253232717514, "rewards/combined_reward/mean": 7.862717723846435, "rewards/combined_reward/std": 2.2790645003318786, "step": 12990 }, { "clip_ratio/high_max": 0.001983695816791927, "clip_ratio/high_mean": 0.00116845135295686, "clip_ratio/low_mean": 0.0008761318562998592, "clip_ratio/low_min": 0.00023580406195833348, "clip_ratio/region_mean": 0.002044583198100251, "completions/clipped_ratio": 0.0, "completions/max_length": 581.6, "completions/max_terminated_length": 581.6, "completions/mean_length": 203.69193267822266, "completions/mean_terminated_length": 203.69193267822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03258673424522082, "epoch": 7.968472604836241, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1332955497.0, "reward": 7.9038537502288815, "reward_std": 0.3381109729409218, "rewards/combined_reward/mean": 7.90385332107544, "rewards/combined_reward/std": 2.4360101342201235, "step": 13020 }, { "clip_ratio/high_max": 0.0018848378286444737, "clip_ratio/high_mean": 0.0010417518688579245, "clip_ratio/low_mean": 0.0009354158955829917, "clip_ratio/low_min": 0.00027110671799164266, "clip_ratio/region_mean": 0.0019771677540120435, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 200.35703887939454, "completions/mean_terminated_length": 200.35703887939454, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03017594600096345, "epoch": 7.986838077747168, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.0244140625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1336025028.0, "reward": 7.971710538864135, "reward_std": 0.2887721508741379, "rewards/combined_reward/mean": 7.97171049118042, "rewards/combined_reward/std": 2.3924246072769164, "step": 13050 }, { "clip_ratio/high_max": 0.001804829020955844, "clip_ratio/high_mean": 0.0010259047010579513, "clip_ratio/low_mean": 0.0007904607391239671, "clip_ratio/low_min": 0.00023857345752089788, "clip_ratio/region_mean": 0.0018163654298643719, "completions/clipped_ratio": 0.0, "completions/max_length": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 202.25364990234374, "completions/mean_terminated_length": 202.25364990234374, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02649456793900396, "epoch": 8.005509641873278, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1339101842.0, "reward": 8.193450355529786, "reward_std": 0.33326144218444825, "rewards/combined_reward/mean": 8.193450212478638, "rewards/combined_reward/std": 2.1087429881095887, "step": 13080 }, { "clip_ratio/high_max": 0.00259352220382425, "clip_ratio/high_mean": 0.0013490869959544701, "clip_ratio/low_mean": 0.000953977212884638, "clip_ratio/low_min": 0.0002682240844781821, "clip_ratio/region_mean": 0.0023030642167820286, "completions/clipped_ratio": 0.0, "completions/max_length": 718.6, "completions/max_terminated_length": 718.6, "completions/mean_length": 199.91745300292968, "completions/mean_terminated_length": 199.91745300292968, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029222187780154247, "epoch": 8.023875114784206, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.007354736328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1342169685.0, "reward": 7.86468243598938, "reward_std": 0.3371377646923065, "rewards/combined_reward/mean": 7.86468243598938, "rewards/combined_reward/std": 2.273501193523407, "step": 13110 }, { "clip_ratio/high_max": 0.002648814419323268, "clip_ratio/high_mean": 0.0015650503051195604, "clip_ratio/low_mean": 0.0009103452454534514, "clip_ratio/low_min": 0.00030044107164333886, "clip_ratio/region_mean": 0.0024753955474200967, "completions/clipped_ratio": 0.0, "completions/max_length": 577.9, "completions/max_terminated_length": 577.9, "completions/mean_length": 210.2916702270508, "completions/mean_terminated_length": 210.2916702270508, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.033611930642897886, "epoch": 8.042240587695133, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1345277365.0, "reward": 8.113387870788575, "reward_std": 0.3027000702917576, "rewards/combined_reward/mean": 8.11338768005371, "rewards/combined_reward/std": 2.3448902368545532, "step": 13140 }, { "clip_ratio/high_max": 0.002586984811447716, "clip_ratio/high_mean": 0.0015676411348977126, "clip_ratio/low_mean": 0.0008936609126976691, "clip_ratio/low_min": 0.000260616846207995, "clip_ratio/region_mean": 0.0024613020301330834, "completions/clipped_ratio": 0.0, "completions/max_length": 672.4, "completions/max_terminated_length": 672.4, "completions/mean_length": 204.9351608276367, "completions/mean_terminated_length": 204.9351608276367, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.03142091403715312, "epoch": 8.06060606060606, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1348364476.0, "reward": 7.892351484298706, "reward_std": 0.41209083795547485, "rewards/combined_reward/mean": 7.892351388931274, "rewards/combined_reward/std": 2.5593346834182737, "step": 13170 }, { "clip_ratio/high_max": 0.0025501299824099988, "clip_ratio/high_mean": 0.0014782689400211287, "clip_ratio/low_mean": 0.0011466142247324267, "clip_ratio/low_min": 0.0004487641148443799, "clip_ratio/region_mean": 0.002624883171908247, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 207.96849517822267, "completions/mean_terminated_length": 207.96849517822267, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.03173814647210141, "epoch": 8.078971533516988, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1351463235.0, "reward": 7.9870563507080075, "reward_std": 0.3077595755457878, "rewards/combined_reward/mean": 7.987056159973145, "rewards/combined_reward/std": 2.339332401752472, "step": 13200 }, { "clip_ratio/high_max": 0.0022360032894842637, "clip_ratio/high_mean": 0.0012444767967584387, "clip_ratio/low_mean": 0.0007386584105309643, "clip_ratio/low_min": 0.00015634113442501983, "clip_ratio/region_mean": 0.001983135201347371, "completions/clipped_ratio": 0.0, "completions/max_length": 571.6, "completions/max_terminated_length": 571.6, "completions/mean_length": 197.70547332763672, "completions/mean_terminated_length": 197.70547332763672, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.02539431316157182, "epoch": 8.097337006427916, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1354522584.0, "reward": 8.111669445037842, "reward_std": 0.3182711601257324, "rewards/combined_reward/mean": 8.111669158935547, "rewards/combined_reward/std": 2.326435649394989, "step": 13230 }, { "clip_ratio/high_max": 0.002252013493368092, "clip_ratio/high_mean": 0.0012931468249007595, "clip_ratio/low_mean": 0.0011168871426586217, "clip_ratio/low_min": 0.0003302406522076732, "clip_ratio/region_mean": 0.0024100339539775933, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 200.75130615234374, "completions/mean_terminated_length": 200.75130615234374, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02919443716915945, "epoch": 8.115702479338843, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.00653076171875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1357593629.0, "reward": 8.239006996154785, "reward_std": 0.32210204005241394, "rewards/combined_reward/mean": 8.239006805419923, "rewards/combined_reward/std": 2.0776790618896483, "step": 13260 }, { "clip_ratio/high_max": 0.00198242645371162, "clip_ratio/high_mean": 0.0011469215186176976, "clip_ratio/low_mean": 0.0008412583961520189, "clip_ratio/low_min": 0.00024702432322859145, "clip_ratio/region_mean": 0.0019881799113742695, "completions/clipped_ratio": 0.0, "completions/max_length": 638.9, "completions/max_terminated_length": 638.9, "completions/mean_length": 216.3151092529297, "completions/mean_terminated_length": 216.3151092529297, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.027155621830994885, "epoch": 8.13406795224977, "frac_reward_zero_std": 0.1583333395421505, "grad_norm": 0.0211181640625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1360724439.0, "reward": 8.347240877151489, "reward_std": 0.3333928033709526, "rewards/combined_reward/mean": 8.347240734100343, "rewards/combined_reward/std": 1.9285810947418214, "step": 13290 }, { "clip_ratio/high_max": 0.0029456002933632894, "clip_ratio/high_mean": 0.0016475408313757118, "clip_ratio/low_mean": 0.0009946716204164356, "clip_ratio/low_min": 0.0002217321440790935, "clip_ratio/region_mean": 0.002642212440211248, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 188.9679733276367, "completions/mean_terminated_length": 188.9679733276367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03500109540764242, "epoch": 8.152433425160698, "frac_reward_zero_std": 0.19166666977107524, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1363750236.0, "reward": 7.997146272659302, "reward_std": 0.3097825452685356, "rewards/combined_reward/mean": 7.9971459865570065, "rewards/combined_reward/std": 2.192604696750641, "step": 13320 }, { "clip_ratio/high_max": 0.0025791969113925007, "clip_ratio/high_mean": 0.001385178572430353, "clip_ratio/low_mean": 0.0010577555947141567, "clip_ratio/low_min": 0.0003336530268522135, "clip_ratio/region_mean": 0.0024429341584133605, "completions/clipped_ratio": 0.0, "completions/max_length": 671.7, "completions/max_terminated_length": 671.7, "completions/mean_length": 200.62500457763673, "completions/mean_terminated_length": 200.62500457763673, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.031402538111433384, "epoch": 8.170798898071626, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 1366820796.0, "reward": 7.8814033508300785, "reward_std": 0.37040064558386804, "rewards/combined_reward/mean": 7.881403112411499, "rewards/combined_reward/std": 2.4807923078536986, "step": 13350 }, { "clip_ratio/high_max": 0.002494035514488739, "clip_ratio/high_mean": 0.001362406496021625, "clip_ratio/low_mean": 0.0008828388614347205, "clip_ratio/low_min": 0.00015733598775113933, "clip_ratio/region_mean": 0.002245245334052015, "completions/clipped_ratio": 0.0, "completions/max_length": 543.4, "completions/max_terminated_length": 543.4, "completions/mean_length": 187.22396392822264, "completions/mean_terminated_length": 187.22396392822264, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028105692216195167, "epoch": 8.189164370982553, "frac_reward_zero_std": 0.20833333693444728, "grad_norm": 0.01446533203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1369839896.0, "reward": 7.8631140232086185, "reward_std": 0.2646182671189308, "rewards/combined_reward/mean": 7.863113737106323, "rewards/combined_reward/std": 2.6214616894721985, "step": 13380 }, { "clip_ratio/high_max": 0.0025699344793489822, "clip_ratio/high_mean": 0.0013132201467669802, "clip_ratio/low_mean": 0.0009640526862009816, "clip_ratio/low_min": 0.0002166912500494315, "clip_ratio/region_mean": 0.002277272860131537, "completions/clipped_ratio": 0.0, "completions/max_length": 612.2, "completions/max_terminated_length": 612.2, "completions/mean_length": 197.63776397705078, "completions/mean_terminated_length": 197.63776397705078, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02902831519022584, "epoch": 8.20752984389348, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.0130615234375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1372898985.0, "reward": 8.111880588531495, "reward_std": 0.31730360686779024, "rewards/combined_reward/mean": 8.111880111694337, "rewards/combined_reward/std": 2.218300831317902, "step": 13410 }, { "clip_ratio/high_max": 0.002810226746078115, "clip_ratio/high_mean": 0.001486737606440632, "clip_ratio/low_mean": 0.0009354200514887149, "clip_ratio/low_min": 0.0002358394587645307, "clip_ratio/region_mean": 0.0024221576473792083, "completions/clipped_ratio": 0.0, "completions/max_length": 538.5, "completions/max_terminated_length": 538.5, "completions/mean_length": 201.95547332763672, "completions/mean_terminated_length": 201.95547332763672, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.028443946751455465, "epoch": 8.225895316804408, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1375974654.0, "reward": 7.976666402816773, "reward_std": 0.26538570821285246, "rewards/combined_reward/mean": 7.9766661643981935, "rewards/combined_reward/std": 2.382442557811737, "step": 13440 }, { "clip_ratio/high_max": 0.0025016367280234894, "clip_ratio/high_mean": 0.0014354312894283793, "clip_ratio/low_mean": 0.0009688517364338622, "clip_ratio/low_min": 0.00028064718984145053, "clip_ratio/region_mean": 0.0024042830087031082, "completions/clipped_ratio": 0.0, "completions/max_length": 655.9, "completions/max_terminated_length": 655.9, "completions/mean_length": 201.93958892822266, "completions/mean_terminated_length": 201.93958892822266, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028922689442212383, "epoch": 8.244260789715335, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.005645751953125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1379050262.0, "reward": 8.057663822174073, "reward_std": 0.32813736945390704, "rewards/combined_reward/mean": 8.057663631439208, "rewards/combined_reward/std": 2.2593481302261353, "step": 13470 }, { "clip_ratio/high_max": 0.0023032822457025757, "clip_ratio/high_mean": 0.0013303479897634437, "clip_ratio/low_mean": 0.0009975073730553656, "clip_ratio/low_min": 0.00029713290593160006, "clip_ratio/region_mean": 0.0023278553616061497, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 197.60391082763672, "completions/mean_terminated_length": 197.60391082763672, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028838758155082665, "epoch": 8.262626262626263, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1382109221.0, "reward": 8.291511344909669, "reward_std": 0.32330141961574554, "rewards/combined_reward/mean": 8.291511297225952, "rewards/combined_reward/std": 1.8988927006721497, "step": 13500 }, { "clip_ratio/high_max": 0.002249551297184856, "clip_ratio/high_mean": 0.0012386288744892227, "clip_ratio/low_mean": 0.0009090218765777536, "clip_ratio/low_min": 0.00023364733836691207, "clip_ratio/region_mean": 0.0021476507376064547, "completions/clipped_ratio": 0.0, "completions/max_length": 597.9, "completions/max_terminated_length": 597.9, "completions/mean_length": 201.71927337646486, "completions/mean_terminated_length": 201.71927337646486, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.02705681447405368, "epoch": 8.28099173553719, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.0062255859375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1385183983.0, "reward": 8.093347263336181, "reward_std": 0.4183322787284851, "rewards/combined_reward/mean": 8.093347120285035, "rewards/combined_reward/std": 2.2907869458198546, "step": 13530 }, { "clip_ratio/high_max": 0.002394853460524852, "clip_ratio/high_mean": 0.0013433062435069588, "clip_ratio/low_mean": 0.0008360527752908335, "clip_ratio/low_min": 0.00021431137841621724, "clip_ratio/region_mean": 0.0021793590334709735, "completions/clipped_ratio": 0.0, "completions/max_length": 612.8, "completions/max_terminated_length": 612.8, "completions/mean_length": 196.04427642822264, "completions/mean_terminated_length": 196.04427642822264, "completions/min_length": 46.5, "completions/min_terminated_length": 46.5, "entropy": 0.02827232420289268, "epoch": 8.299357208448118, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.005157470703125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1388236953.0, "reward": 8.401779747009277, "reward_std": 0.3539200112223625, "rewards/combined_reward/mean": 8.401779651641846, "rewards/combined_reward/std": 1.8270691990852357, "step": 13560 }, { "clip_ratio/high_max": 0.0025535100401612, "clip_ratio/high_mean": 0.0014393254853833544, "clip_ratio/low_mean": 0.0009589583478373243, "clip_ratio/low_min": 0.0002463948830457715, "clip_ratio/region_mean": 0.0023982838164859762, "completions/clipped_ratio": 0.0, "completions/max_length": 643.2, "completions/max_terminated_length": 643.2, "completions/mean_length": 201.3229187011719, "completions/mean_terminated_length": 201.3229187011719, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02900051500182599, "epoch": 8.317722681359045, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1391310193.0, "reward": 7.933571147918701, "reward_std": 0.352197590470314, "rewards/combined_reward/mean": 7.933570909500122, "rewards/combined_reward/std": 2.3801127791404726, "step": 13590 }, { "clip_ratio/high_max": 0.0026124562646145932, "clip_ratio/high_mean": 0.001478564903906469, "clip_ratio/low_mean": 0.000907983947035973, "clip_ratio/low_min": 0.00020647750231243359, "clip_ratio/region_mean": 0.0023865488527614314, "completions/clipped_ratio": 0.0, "completions/max_length": 588.0, "completions/max_terminated_length": 588.0, "completions/mean_length": 204.6007843017578, "completions/mean_terminated_length": 204.6007843017578, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.02872603917494416, "epoch": 8.336088154269973, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.0067138671875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1394396020.0, "reward": 7.882661962509156, "reward_std": 0.3411796674132347, "rewards/combined_reward/mean": 7.88266167640686, "rewards/combined_reward/std": 2.5364654302597045, "step": 13620 }, { "clip_ratio/high_max": 0.002373699229307628, "clip_ratio/high_mean": 0.001300079026259482, "clip_ratio/low_mean": 0.001080662834768494, "clip_ratio/low_min": 0.0003989285129743318, "clip_ratio/region_mean": 0.0023807418649084864, "completions/clipped_ratio": 0.0, "completions/max_length": 568.2, "completions/max_terminated_length": 568.2, "completions/mean_length": 199.50521545410157, "completions/mean_terminated_length": 199.50521545410157, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.03100878872598211, "epoch": 8.3544536271809, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1397462280.0, "reward": 8.068643140792847, "reward_std": 0.33193385452032087, "rewards/combined_reward/mean": 8.068642711639404, "rewards/combined_reward/std": 2.3087647438049315, "step": 13650 }, { "clip_ratio/high_max": 0.0029619159402500374, "clip_ratio/high_mean": 0.001644368791554977, "clip_ratio/low_mean": 0.000991670650182641, "clip_ratio/low_min": 0.0003123783668949424, "clip_ratio/region_mean": 0.002636039450104969, "completions/clipped_ratio": 0.0, "completions/max_length": 701.3, "completions/max_terminated_length": 701.3, "completions/mean_length": 201.58594207763673, "completions/mean_terminated_length": 201.58594207763673, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.03179114543211957, "epoch": 8.372819100091828, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0284423828125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1400536530.0, "reward": 8.396274518966674, "reward_std": 0.3641042083501816, "rewards/combined_reward/mean": 8.39627456665039, "rewards/combined_reward/std": 1.8568087458610534, "step": 13680 }, { "clip_ratio/high_max": 0.002182144906449442, "clip_ratio/high_mean": 0.0012568903730425518, "clip_ratio/low_mean": 0.0008384729220779263, "clip_ratio/low_min": 0.00019723644824504543, "clip_ratio/region_mean": 0.0020953632886327495, "completions/clipped_ratio": 0.0, "completions/max_length": 625.5, "completions/max_terminated_length": 625.5, "completions/mean_length": 200.8651092529297, "completions/mean_terminated_length": 200.8651092529297, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.028151536375905076, "epoch": 8.391184573002755, "frac_reward_zero_std": 0.22916667461395263, "grad_norm": 0.02880859375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1403608012.0, "reward": 7.8734218120574955, "reward_std": 0.34201278239488603, "rewards/combined_reward/mean": 7.873421478271484, "rewards/combined_reward/std": 2.453501307964325, "step": 13710 }, { "clip_ratio/high_max": 0.003004668278056973, "clip_ratio/high_mean": 0.0016639730796062697, "clip_ratio/low_mean": 0.0010986470131683745, "clip_ratio/low_min": 0.00035944703389153193, "clip_ratio/region_mean": 0.0027626200870145114, "completions/clipped_ratio": 0.0, "completions/max_length": 612.4, "completions/max_terminated_length": 612.4, "completions/mean_length": 202.4635482788086, "completions/mean_terminated_length": 202.4635482788086, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03250249706519147, "epoch": 8.409550045913683, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1406685632.0, "reward": 7.920281934738159, "reward_std": 0.3315482273697853, "rewards/combined_reward/mean": 7.9202817440032955, "rewards/combined_reward/std": 2.3207436084747313, "step": 13740 }, { "clip_ratio/high_max": 0.0021629405848216264, "clip_ratio/high_mean": 0.0012192891729985906, "clip_ratio/low_mean": 0.0007698515461621961, "clip_ratio/low_min": 0.00016933960068854503, "clip_ratio/region_mean": 0.0019891407175843295, "completions/clipped_ratio": 0.0, "completions/max_length": 532.6, "completions/max_terminated_length": 532.6, "completions/mean_length": 199.32656860351562, "completions/mean_terminated_length": 199.32656860351562, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.027215650506938496, "epoch": 8.42791551882461, "frac_reward_zero_std": 0.13333333730697633, "grad_norm": 0.0035858154296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1409751206.0, "reward": 8.12939395904541, "reward_std": 0.36719831079244614, "rewards/combined_reward/mean": 8.129393815994263, "rewards/combined_reward/std": 2.2580177664756773, "step": 13770 }, { "clip_ratio/high_max": 0.0020593486241220187, "clip_ratio/high_mean": 0.0011170251230699553, "clip_ratio/low_mean": 0.000809434951285463, "clip_ratio/low_min": 0.00020239542282070034, "clip_ratio/region_mean": 0.0019264600938186049, "completions/clipped_ratio": 0.0, "completions/max_length": 575.4, "completions/max_terminated_length": 575.4, "completions/mean_length": 192.16380615234374, "completions/mean_terminated_length": 192.16380615234374, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02499668227198223, "epoch": 8.446280991735538, "frac_reward_zero_std": 0.20833333991467953, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1412789275.0, "reward": 8.220582485198975, "reward_std": 0.24562513828277588, "rewards/combined_reward/mean": 8.220582008361816, "rewards/combined_reward/std": 2.1804901480674745, "step": 13800 }, { "clip_ratio/high_max": 0.0023037959736636063, "clip_ratio/high_mean": 0.0013270635813872407, "clip_ratio/low_mean": 0.0007866930486367589, "clip_ratio/low_min": 0.00018167543724606123, "clip_ratio/region_mean": 0.00211375664997225, "completions/clipped_ratio": 0.0, "completions/max_length": 573.1, "completions/max_terminated_length": 573.1, "completions/mean_length": 202.2088592529297, "completions/mean_terminated_length": 202.2088592529297, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.028999697013447682, "epoch": 8.464646464646465, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 1415865917.0, "reward": 8.0481746673584, "reward_std": 0.40375624001026156, "rewards/combined_reward/mean": 8.048174476623535, "rewards/combined_reward/std": 2.4106324911117554, "step": 13830 }, { "clip_ratio/high_max": 0.0024348198348889127, "clip_ratio/high_mean": 0.0014251227944138615, "clip_ratio/low_mean": 0.0009340562127666392, "clip_ratio/low_min": 0.00031594999042378427, "clip_ratio/region_mean": 0.0023591790202772247, "completions/clipped_ratio": 0.0, "completions/max_length": 554.2, "completions/max_terminated_length": 554.2, "completions/mean_length": 202.8375045776367, "completions/mean_terminated_length": 202.8375045776367, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.027729340045092008, "epoch": 8.483011937557393, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1418944973.0, "reward": 8.089618921279907, "reward_std": 0.3474583253264427, "rewards/combined_reward/mean": 8.089618825912476, "rewards/combined_reward/std": 2.154957127571106, "step": 13860 }, { "clip_ratio/high_max": 0.0020433665362361354, "clip_ratio/high_mean": 0.001117626620058824, "clip_ratio/low_mean": 0.0007405084402610858, "clip_ratio/low_min": 0.0001835225360991899, "clip_ratio/region_mean": 0.001858135069293591, "completions/clipped_ratio": 0.0, "completions/max_length": 517.2, "completions/max_terminated_length": 517.2, "completions/mean_length": 195.70859985351564, "completions/mean_terminated_length": 195.70859985351564, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "entropy": 0.025359725820211072, "epoch": 8.501377410468319, "frac_reward_zero_std": 0.20416667200624944, "grad_norm": 0.01507568359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1421996654.0, "reward": 7.945265865325927, "reward_std": 0.41903025209903716, "rewards/combined_reward/mean": 7.945265579223633, "rewards/combined_reward/std": 2.515912914276123, "step": 13890 }, { "clip_ratio/high_max": 0.0021530446984494725, "clip_ratio/high_mean": 0.001215534507476453, "clip_ratio/low_mean": 0.0009261511098884512, "clip_ratio/low_min": 0.0002976643059810158, "clip_ratio/region_mean": 0.0021416856220942766, "completions/clipped_ratio": 0.0, "completions/max_length": 530.5, "completions/max_terminated_length": 530.5, "completions/mean_length": 205.27734985351563, "completions/mean_terminated_length": 205.27734985351563, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.02649524201018115, "epoch": 8.519742883379248, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1425085079.0, "reward": 8.044724798202514, "reward_std": 0.3201822072267532, "rewards/combined_reward/mean": 8.044724559783935, "rewards/combined_reward/std": 2.117205023765564, "step": 13920 }, { "clip_ratio/high_max": 0.0023012666148133575, "clip_ratio/high_mean": 0.0014076585949320966, "clip_ratio/low_mean": 0.0008476537183014443, "clip_ratio/low_min": 0.0002972087114661311, "clip_ratio/region_mean": 0.002255312310201892, "completions/clipped_ratio": 0.0, "completions/max_length": 671.9, "completions/max_terminated_length": 671.9, "completions/mean_length": 206.28490142822267, "completions/mean_terminated_length": 206.28490142822267, "completions/min_length": 47.0, "completions/min_terminated_length": 47.0, "entropy": 0.028439629698793094, "epoch": 8.538108356290174, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.006317138671875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1428177373.0, "reward": 8.070912599563599, "reward_std": 0.27675081193447115, "rewards/combined_reward/mean": 8.070912218093872, "rewards/combined_reward/std": 2.4064706563949585, "step": 13950 }, { "clip_ratio/high_max": 0.002495237386513812, "clip_ratio/high_mean": 0.0013405020642191328, "clip_ratio/low_mean": 0.0010713250174982627, "clip_ratio/low_min": 0.000331109522570235, "clip_ratio/region_mean": 0.0024118270705609274, "completions/clipped_ratio": 0.0, "completions/max_length": 603.7, "completions/max_terminated_length": 603.7, "completions/mean_length": 195.6476608276367, "completions/mean_terminated_length": 195.6476608276367, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03093447753538688, "epoch": 8.556473829201101, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1431228820.0, "reward": 7.992555332183838, "reward_std": 0.3323943018913269, "rewards/combined_reward/mean": 7.992555046081543, "rewards/combined_reward/std": 2.4151178002357483, "step": 13980 }, { "clip_ratio/high_max": 0.002586718254800265, "clip_ratio/high_mean": 0.0015587752372084652, "clip_ratio/low_mean": 0.0010373291640765577, "clip_ratio/low_min": 0.0003178097019069052, "clip_ratio/region_mean": 0.0025961043895222245, "completions/clipped_ratio": 0.0, "completions/max_length": 542.5, "completions/max_terminated_length": 542.5, "completions/mean_length": 191.69245452880858, "completions/mean_terminated_length": 191.69245452880858, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02875762494901816, "epoch": 8.574839302112029, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1434265079.0, "reward": 8.074105310440064, "reward_std": 0.31561674624681474, "rewards/combined_reward/mean": 8.07410488128662, "rewards/combined_reward/std": 2.2865672945976256, "step": 14010 }, { "clip_ratio/high_max": 0.003017870203863519, "clip_ratio/high_mean": 0.001618112022212396, "clip_ratio/low_mean": 0.000978420119645307, "clip_ratio/low_min": 0.0002971201225591358, "clip_ratio/region_mean": 0.002596532128518447, "completions/clipped_ratio": 0.0, "completions/max_length": 513.8, "completions/max_terminated_length": 513.8, "completions/mean_length": 192.53359832763672, "completions/mean_terminated_length": 192.53359832763672, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030991723015904426, "epoch": 8.593204775022956, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1437304568.0, "reward": 7.932095241546631, "reward_std": 0.3214726343750954, "rewards/combined_reward/mean": 7.932095003128052, "rewards/combined_reward/std": 2.4626341581344606, "step": 14040 }, { "clip_ratio/high_max": 0.002791183771235713, "clip_ratio/high_mean": 0.0016124208685747968, "clip_ratio/low_mean": 0.0009151812447574533, "clip_ratio/low_min": 0.00021983805563650095, "clip_ratio/region_mean": 0.002527602112968452, "completions/clipped_ratio": 0.0, "completions/max_length": 556.9, "completions/max_terminated_length": 556.9, "completions/mean_length": 200.1315139770508, "completions/mean_terminated_length": 200.1315139770508, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.030871748582770427, "epoch": 8.611570247933884, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1440373233.0, "reward": 7.992812204360962, "reward_std": 0.3865604132413864, "rewards/combined_reward/mean": 7.992811775207519, "rewards/combined_reward/std": 2.175676369667053, "step": 14070 }, { "clip_ratio/high_max": 0.0023231251868613373, "clip_ratio/high_mean": 0.0014478339258251556, "clip_ratio/low_mean": 0.0010359490472183098, "clip_ratio/low_min": 0.000261576980362103, "clip_ratio/region_mean": 0.0024837829875953804, "completions/clipped_ratio": 0.0, "completions/max_length": 584.5, "completions/max_terminated_length": 584.5, "completions/mean_length": 199.3026107788086, "completions/mean_terminated_length": 199.3026107788086, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03319857084813217, "epoch": 8.629935720844811, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.03125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1443438715.0, "reward": 7.973708391189575, "reward_std": 0.38130598068237304, "rewards/combined_reward/mean": 7.9737081050872805, "rewards/combined_reward/std": 2.279932975769043, "step": 14100 }, { "clip_ratio/high_max": 0.002540518336657745, "clip_ratio/high_mean": 0.0014285228156950326, "clip_ratio/low_mean": 0.0011392754150923187, "clip_ratio/low_min": 0.000259702533124558, "clip_ratio/region_mean": 0.0025677982286045637, "completions/clipped_ratio": 0.0, "completions/max_length": 548.5, "completions/max_terminated_length": 548.5, "completions/mean_length": 197.05912017822266, "completions/mean_terminated_length": 197.05912017822266, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03166913399472833, "epoch": 8.648301193755739, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1446495582.0, "reward": 8.220362091064453, "reward_std": 0.297977077960968, "rewards/combined_reward/mean": 8.220362329483033, "rewards/combined_reward/std": 2.140598440170288, "step": 14130 }, { "clip_ratio/high_max": 0.002604841145997246, "clip_ratio/high_mean": 0.0015001592445817854, "clip_ratio/low_mean": 0.0011908067637705245, "clip_ratio/low_min": 0.0004607837698131334, "clip_ratio/region_mean": 0.0026909659830077243, "completions/clipped_ratio": 0.0, "completions/max_length": 585.7, "completions/max_terminated_length": 585.7, "completions/mean_length": 197.65469055175782, "completions/mean_terminated_length": 197.65469055175782, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03496728898026049, "epoch": 8.666666666666666, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.00775146484375, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 1449554736.0, "reward": 7.762375783920288, "reward_std": 0.4174013495445251, "rewards/combined_reward/mean": 7.762375688552856, "rewards/combined_reward/std": 2.3687588572502136, "step": 14160 }, { "clip_ratio/high_max": 0.0022882518567106064, "clip_ratio/high_mean": 0.0013143055539937146, "clip_ratio/low_mean": 0.0010601107034744928, "clip_ratio/low_min": 0.00021536421603135142, "clip_ratio/region_mean": 0.0023744162511623776, "completions/clipped_ratio": 0.0, "completions/max_length": 555.4, "completions/max_terminated_length": 555.4, "completions/mean_length": 198.38281860351563, "completions/mean_terminated_length": 198.38281860351563, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.03016394772566855, "epoch": 8.685032139577594, "frac_reward_zero_std": 0.19583333916962148, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1452616686.0, "reward": 8.14574236869812, "reward_std": 0.3326534077525139, "rewards/combined_reward/mean": 8.145742273330688, "rewards/combined_reward/std": 2.163506293296814, "step": 14190 }, { "clip_ratio/high_max": 0.0022626981410818797, "clip_ratio/high_mean": 0.0012344461338216206, "clip_ratio/low_mean": 0.0009665369096789315, "clip_ratio/low_min": 0.00024908964987844227, "clip_ratio/region_mean": 0.0022009830516860045, "completions/clipped_ratio": 0.0, "completions/max_length": 559.4, "completions/max_terminated_length": 559.4, "completions/mean_length": 194.03802490234375, "completions/mean_terminated_length": 194.03802490234375, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028655529061021905, "epoch": 8.703397612488521, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.044677734375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1455661952.0, "reward": 7.969976568222046, "reward_std": 0.4065878793597221, "rewards/combined_reward/mean": 7.969976377487183, "rewards/combined_reward/std": 2.4533603310585024, "step": 14220 }, { "clip_ratio/high_max": 0.0023721541234408504, "clip_ratio/high_mean": 0.0013503207570465746, "clip_ratio/low_mean": 0.000994242032902548, "clip_ratio/low_min": 0.0002638624154011874, "clip_ratio/region_mean": 0.0023445627761248034, "completions/clipped_ratio": 0.0, "completions/max_length": 579.4, "completions/max_terminated_length": 579.4, "completions/mean_length": 200.48516235351562, "completions/mean_terminated_length": 200.48516235351562, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0306171293836087, "epoch": 8.721763085399449, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1458731975.0, "reward": 7.9792266368865965, "reward_std": 0.33721729815006257, "rewards/combined_reward/mean": 7.97922682762146, "rewards/combined_reward/std": 2.3062962174415587, "step": 14250 }, { "clip_ratio/high_max": 0.002712161422338492, "clip_ratio/high_mean": 0.0015218851971440018, "clip_ratio/low_mean": 0.0010526224757389476, "clip_ratio/low_min": 0.00028025394349242563, "clip_ratio/region_mean": 0.002574507679673843, "completions/clipped_ratio": 0.0, "completions/max_length": 634.1, "completions/max_terminated_length": 634.1, "completions/mean_length": 197.9937530517578, "completions/mean_terminated_length": 197.9937530517578, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03200047250526647, "epoch": 8.740128558310376, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 1461792431.0, "reward": 7.775072717666626, "reward_std": 0.34720529019832613, "rewards/combined_reward/mean": 7.775072479248047, "rewards/combined_reward/std": 2.442213100194931, "step": 14280 }, { "clip_ratio/high_max": 0.0028622944048644666, "clip_ratio/high_mean": 0.0015882690677244682, "clip_ratio/low_mean": 0.0010212779337355944, "clip_ratio/low_min": 0.00025371867644328935, "clip_ratio/region_mean": 0.002609546994790435, "completions/clipped_ratio": 0.0, "completions/max_length": 621.2, "completions/max_terminated_length": 621.2, "completions/mean_length": 203.81693267822266, "completions/mean_terminated_length": 203.81693267822266, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03294820158431928, "epoch": 8.758494031221304, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1464875248.0, "reward": 7.94718599319458, "reward_std": 0.3262374669313431, "rewards/combined_reward/mean": 7.9471855640411375, "rewards/combined_reward/std": 2.3759200870990753, "step": 14310 }, { "clip_ratio/high_max": 0.002194818376059023, "clip_ratio/high_mean": 0.0012934855801480202, "clip_ratio/low_mean": 0.0006418806963968867, "clip_ratio/low_min": 0.00011229130153272611, "clip_ratio/region_mean": 0.001935366270481609, "completions/clipped_ratio": 0.0, "completions/max_length": 689.8, "completions/max_terminated_length": 689.8, "completions/mean_length": 200.39974365234374, "completions/mean_terminated_length": 200.39974365234374, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028349976645161708, "epoch": 8.776859504132231, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1467944943.0, "reward": 8.101503372192383, "reward_std": 0.36437528729438784, "rewards/combined_reward/mean": 8.101503133773804, "rewards/combined_reward/std": 2.272334122657776, "step": 14340 }, { "clip_ratio/high_max": 0.0022583056236423244, "clip_ratio/high_mean": 0.001332976464376164, "clip_ratio/low_mean": 0.0011236067928014867, "clip_ratio/low_min": 0.0003840033986004225, "clip_ratio/region_mean": 0.0024565832611794275, "completions/clipped_ratio": 0.0, "completions/max_length": 612.5, "completions/max_terminated_length": 612.5, "completions/mean_length": 201.7268295288086, "completions/mean_terminated_length": 201.7268295288086, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.030313435088222224, "epoch": 8.795224977043159, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1471019734.0, "reward": 7.867444753646851, "reward_std": 0.39743279218673705, "rewards/combined_reward/mean": 7.8674445152282715, "rewards/combined_reward/std": 2.3347797632217406, "step": 14370 }, { "clip_ratio/high_max": 0.002581215579508959, "clip_ratio/high_mean": 0.0014569224691513226, "clip_ratio/low_mean": 0.0009687114389331934, "clip_ratio/low_min": 0.00020245174406833635, "clip_ratio/region_mean": 0.002425633918270857, "completions/clipped_ratio": 0.0, "completions/max_length": 595.3, "completions/max_terminated_length": 595.3, "completions/mean_length": 202.83177490234374, "completions/mean_terminated_length": 202.83177490234374, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.031120290715868273, "epoch": 8.813590449954086, "frac_reward_zero_std": 0.12500000186264515, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1474098768.0, "reward": 8.095383024215698, "reward_std": 0.397204315662384, "rewards/combined_reward/mean": 8.095382690429688, "rewards/combined_reward/std": 2.141866147518158, "step": 14400 }, { "clip_ratio/high_max": 0.002289884134855432, "clip_ratio/high_mean": 0.0013172583438669486, "clip_ratio/low_mean": 0.0008320012614907076, "clip_ratio/low_min": 0.000238216433111423, "clip_ratio/region_mean": 0.002149259591048273, "completions/clipped_ratio": 0.0, "completions/max_length": 506.5, "completions/max_terminated_length": 506.5, "completions/mean_length": 195.64662017822266, "completions/mean_terminated_length": 195.64662017822266, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03241090518422425, "epoch": 8.831343740434649, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1477150211.0, "reward": 8.00532865524292, "reward_std": 0.34465041756629944, "rewards/combined_reward/mean": 8.005328512191772, "rewards/combined_reward/std": 2.291551423072815, "step": 14430 }, { "clip_ratio/high_max": 0.0017101697022250542, "clip_ratio/high_mean": 0.00092777434377543, "clip_ratio/low_mean": 0.0006975480208590549, "clip_ratio/low_min": 0.0001757036996423267, "clip_ratio/region_mean": 0.0016253223763972831, "completions/clipped_ratio": 0.0, "completions/max_length": 681.4, "completions/max_terminated_length": 681.4, "completions/mean_length": 195.90208892822267, "completions/mean_terminated_length": 195.90208892822267, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.025838980249439678, "epoch": 8.849709213345577, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1480202635.0, "reward": 8.147314739227294, "reward_std": 0.3121831275522709, "rewards/combined_reward/mean": 8.147314548492432, "rewards/combined_reward/std": 2.23541841506958, "step": 14460 }, { "clip_ratio/high_max": 0.0019174896893673576, "clip_ratio/high_mean": 0.0011181337791640543, "clip_ratio/low_mean": 0.0006981950241121619, "clip_ratio/low_min": 0.0002199178394221235, "clip_ratio/region_mean": 0.0018163287886030352, "completions/clipped_ratio": 0.0, "completions/max_length": 564.8, "completions/max_terminated_length": 564.8, "completions/mean_length": 201.97552642822265, "completions/mean_terminated_length": 201.97552642822265, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03196623822053273, "epoch": 8.868074686256504, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01507568359375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1483278381.0, "reward": 8.371120119094849, "reward_std": 0.3258804127573967, "rewards/combined_reward/mean": 8.371119928359985, "rewards/combined_reward/std": 1.9000550150871276, "step": 14490 }, { "clip_ratio/high_max": 0.002530006857220239, "clip_ratio/high_mean": 0.0013420968568728616, "clip_ratio/low_mean": 0.0007547824603534537, "clip_ratio/low_min": 0.00018779599098100636, "clip_ratio/region_mean": 0.0020968792969749, "completions/clipped_ratio": 0.0, "completions/max_length": 536.8, "completions/max_terminated_length": 536.8, "completions/mean_length": 187.9875061035156, "completions/mean_terminated_length": 187.9875061035156, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03273457541751365, "epoch": 8.886440159167432, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.015380859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1486300413.0, "reward": 8.010940599441529, "reward_std": 0.36791117042303084, "rewards/combined_reward/mean": 8.010940027236938, "rewards/combined_reward/std": 2.333020329475403, "step": 14520 }, { "clip_ratio/high_max": 0.0020805302629014477, "clip_ratio/high_mean": 0.0011455471882072742, "clip_ratio/low_mean": 0.000821557808618915, "clip_ratio/low_min": 0.00031302218873558254, "clip_ratio/region_mean": 0.0019671050084677216, "completions/clipped_ratio": 0.0, "completions/max_length": 602.9, "completions/max_terminated_length": 602.9, "completions/mean_length": 208.29766082763672, "completions/mean_terminated_length": 208.29766082763672, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.029626706327932576, "epoch": 8.904805632078359, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1489400436.0, "reward": 8.194556045532227, "reward_std": 0.35660928040742873, "rewards/combined_reward/mean": 8.194555807113648, "rewards/combined_reward/std": 2.0784597396850586, "step": 14550 }, { "clip_ratio/high_max": 0.0019363833217842816, "clip_ratio/high_mean": 0.0011247892250442722, "clip_ratio/low_mean": 0.00074250680457529, "clip_ratio/low_min": 0.0002304333400388714, "clip_ratio/region_mean": 0.0018672960053663702, "completions/clipped_ratio": 0.0, "completions/max_length": 555.5, "completions/max_terminated_length": 555.5, "completions/mean_length": 199.55859832763673, "completions/mean_terminated_length": 199.55859832763673, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.030490923851417998, "epoch": 8.923171104989287, "frac_reward_zero_std": 0.21250000447034836, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1492466901.0, "reward": 7.959870719909668, "reward_std": 0.3616245940327644, "rewards/combined_reward/mean": 7.959870338439941, "rewards/combined_reward/std": 2.4610998153686525, "step": 14580 }, { "clip_ratio/high_max": 0.0017562955913793606, "clip_ratio/high_mean": 0.0010802869125958144, "clip_ratio/low_mean": 0.0007950508090289077, "clip_ratio/low_min": 0.0001483338494532897, "clip_ratio/region_mean": 0.0018753377386019564, "completions/clipped_ratio": 0.0, "completions/max_length": 512.2, "completions/max_terminated_length": 512.2, "completions/mean_length": 182.7791717529297, "completions/mean_terminated_length": 182.7791717529297, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027839546386773387, "epoch": 8.941536577900214, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1495468933.0, "reward": 7.780278539657592, "reward_std": 0.28239518105983735, "rewards/combined_reward/mean": 7.780278110504151, "rewards/combined_reward/std": 2.4564106941223143, "step": 14610 }, { "clip_ratio/high_max": 0.001770266033418011, "clip_ratio/high_mean": 0.0010018818711008256, "clip_ratio/low_mean": 0.0008697020195540972, "clip_ratio/low_min": 0.000275570619123755, "clip_ratio/region_mean": 0.0018715838828939012, "completions/clipped_ratio": 0.0, "completions/max_length": 592.4, "completions/max_terminated_length": 592.4, "completions/mean_length": 201.93594055175782, "completions/mean_terminated_length": 201.93594055175782, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027303885792692502, "epoch": 8.959902050811142, "frac_reward_zero_std": 0.1250000037252903, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1498544527.0, "reward": 8.064399766921998, "reward_std": 0.33972382098436354, "rewards/combined_reward/mean": 8.064399433135986, "rewards/combined_reward/std": 2.122112047672272, "step": 14640 }, { "clip_ratio/high_max": 0.0018196965834552732, "clip_ratio/high_mean": 0.0010638713947021946, "clip_ratio/low_mean": 0.0006924109295747864, "clip_ratio/low_min": 0.00014076335889209683, "clip_ratio/region_mean": 0.001756282318092417, "completions/clipped_ratio": 0.0, "completions/max_length": 518.7, "completions/max_terminated_length": 518.7, "completions/mean_length": 195.91510772705078, "completions/mean_terminated_length": 195.91510772705078, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.030078660355259974, "epoch": 8.978267523722069, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1501597001.0, "reward": 7.803607368469239, "reward_std": 0.33298497647047043, "rewards/combined_reward/mean": 7.8036072731018065, "rewards/combined_reward/std": 2.572816479206085, "step": 14670 }, { "clip_ratio/high_max": 0.001996932267017352, "clip_ratio/high_mean": 0.001141747617657529, "clip_ratio/low_mean": 0.0008645302686394037, "clip_ratio/low_min": 0.0003422163485083729, "clip_ratio/region_mean": 0.002006277887024529, "completions/clipped_ratio": 0.0, "completions/max_length": 625.3, "completions/max_terminated_length": 625.3, "completions/mean_length": 209.97891235351562, "completions/mean_terminated_length": 209.97891235351562, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02857555525843054, "epoch": 8.996632996632997, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1504703480.0, "reward": 7.9838926792144775, "reward_std": 0.33192437812685965, "rewards/combined_reward/mean": 7.983892583847046, "rewards/combined_reward/std": 2.357188415527344, "step": 14700 }, { "clip_ratio/high_max": 0.002607998736851589, "clip_ratio/high_mean": 0.0014716075369744699, "clip_ratio/low_mean": 0.0010431988214163613, "clip_ratio/low_min": 0.00029898464243637674, "clip_ratio/region_mean": 0.0025148063694836846, "completions/clipped_ratio": 0.0, "completions/max_length": 726.1818181818181, "completions/max_terminated_length": 726.1818181818181, "completions/mean_length": 203.05634932084516, "completions/mean_terminated_length": 203.05634932084516, "completions/min_length": 37.81818181818182, "completions/min_terminated_length": 37.81818181818182, "entropy": 0.030650700106606133, "epoch": 9.015304560759105, "frac_reward_zero_std": 0.16287879103964026, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1508091366.0, "reward": 8.128058043393223, "reward_std": 0.28992348367517645, "rewards/combined_reward/mean": 8.128057956695557, "rewards/combined_reward/std": 2.0097497810016978, "step": 14730 }, { "clip_ratio/high_max": 0.0023288747475210887, "clip_ratio/high_mean": 0.0012773628718908488, "clip_ratio/low_mean": 0.0009162967315205605, "clip_ratio/low_min": 0.0002824761795636732, "clip_ratio/region_mean": 0.002193659612870154, "completions/clipped_ratio": 0.0, "completions/max_length": 525.5, "completions/max_terminated_length": 525.5, "completions/mean_length": 208.01641082763672, "completions/mean_terminated_length": 208.01641082763672, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.02934290051149825, "epoch": 9.033670033670033, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1511190309.0, "reward": 7.917802619934082, "reward_std": 0.349201063811779, "rewards/combined_reward/mean": 7.917802381515503, "rewards/combined_reward/std": 2.572335720062256, "step": 14760 }, { "clip_ratio/high_max": 0.0025712643017565522, "clip_ratio/high_mean": 0.0013950059807636233, "clip_ratio/low_mean": 0.0010144926424497196, "clip_ratio/low_min": 0.0002952759386971593, "clip_ratio/region_mean": 0.002409498591926725, "completions/clipped_ratio": 0.0, "completions/max_length": 581.4, "completions/max_terminated_length": 581.4, "completions/mean_length": 203.48281860351562, "completions/mean_terminated_length": 203.48281860351562, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03044368727132678, "epoch": 9.05203550658096, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1514271843.0, "reward": 8.047529268264771, "reward_std": 0.346893772482872, "rewards/combined_reward/mean": 8.047529125213623, "rewards/combined_reward/std": 2.3243228435516357, "step": 14790 }, { "clip_ratio/high_max": 0.002438944930327125, "clip_ratio/high_mean": 0.0013687382898448655, "clip_ratio/low_mean": 0.001090784168142515, "clip_ratio/low_min": 0.00028379290451994164, "clip_ratio/region_mean": 0.0024595224482861036, "completions/clipped_ratio": 0.0, "completions/max_length": 618.3, "completions/max_terminated_length": 618.3, "completions/mean_length": 206.25052642822266, "completions/mean_terminated_length": 206.25052642822266, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.03082168713832895, "epoch": 9.070400979491888, "frac_reward_zero_std": 0.12916667051613331, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1517364005.0, "reward": 8.087829494476319, "reward_std": 0.3072432100772858, "rewards/combined_reward/mean": 8.087829113006592, "rewards/combined_reward/std": 2.158422887325287, "step": 14820 }, { "clip_ratio/high_max": 0.0024161235754339336, "clip_ratio/high_mean": 0.0013142209686823965, "clip_ratio/low_mean": 0.0009657289750369576, "clip_ratio/low_min": 0.00030888356389671874, "clip_ratio/region_mean": 0.002279949928439843, "completions/clipped_ratio": 0.0, "completions/max_length": 533.5, "completions/max_terminated_length": 533.5, "completions/mean_length": 201.9737060546875, "completions/mean_terminated_length": 201.9737060546875, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.029427714894215265, "epoch": 9.088766452402815, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0103759765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1520439744.0, "reward": 7.91480450630188, "reward_std": 0.33274187296628954, "rewards/combined_reward/mean": 7.914804172515869, "rewards/combined_reward/std": 2.517943501472473, "step": 14850 }, { "clip_ratio/high_max": 0.0022771007444437903, "clip_ratio/high_mean": 0.0013396964398756002, "clip_ratio/low_mean": 0.0007401828001699566, "clip_ratio/low_min": 0.00015850282191725757, "clip_ratio/region_mean": 0.002079879246108855, "completions/clipped_ratio": 0.0, "completions/max_length": 563.6, "completions/max_terminated_length": 563.6, "completions/mean_length": 195.41120300292968, "completions/mean_terminated_length": 195.41120300292968, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.02617021972934405, "epoch": 9.107131925313743, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1523490283.0, "reward": 8.17085795402527, "reward_std": 0.4142463684082031, "rewards/combined_reward/mean": 8.170857477188111, "rewards/combined_reward/std": 2.1790772438049317, "step": 14880 }, { "clip_ratio/high_max": 0.002178663477631441, "clip_ratio/high_mean": 0.0012525087498943322, "clip_ratio/low_mean": 0.0009113890203783134, "clip_ratio/low_min": 0.0003299551846187872, "clip_ratio/region_mean": 0.0021638977617840283, "completions/clipped_ratio": 0.0, "completions/max_length": 616.3, "completions/max_terminated_length": 616.3, "completions/mean_length": 203.87266235351564, "completions/mean_terminated_length": 203.87266235351564, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.02696800829532246, "epoch": 9.12549739822467, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1526573314.0, "reward": 8.334821414947509, "reward_std": 0.3099849671125412, "rewards/combined_reward/mean": 8.334821081161499, "rewards/combined_reward/std": 2.033106815814972, "step": 14910 }, { "clip_ratio/high_max": 0.0031349723450451467, "clip_ratio/high_mean": 0.0017979529463142778, "clip_ratio/low_mean": 0.0008981460390108017, "clip_ratio/low_min": 0.0002485879970966683, "clip_ratio/region_mean": 0.0026960989790192494, "completions/clipped_ratio": 0.0, "completions/max_length": 724.8, "completions/max_terminated_length": 724.8, "completions/mean_length": 199.5572982788086, "completions/mean_terminated_length": 199.5572982788086, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03250728769538303, "epoch": 9.143862871135598, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1529639774.0, "reward": 8.11400728225708, "reward_std": 0.31883329153060913, "rewards/combined_reward/mean": 8.114006900787354, "rewards/combined_reward/std": 2.133583104610443, "step": 14940 }, { "clip_ratio/high_max": 0.0024404707461750754, "clip_ratio/high_mean": 0.0014339591038151412, "clip_ratio/low_mean": 0.0009237152908705563, "clip_ratio/low_min": 0.0002824885646987241, "clip_ratio/region_mean": 0.002357674392987974, "completions/clipped_ratio": 0.0, "completions/max_length": 686.5, "completions/max_terminated_length": 686.5, "completions/mean_length": 204.34948272705077, "completions/mean_terminated_length": 204.34948272705077, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03346719796148439, "epoch": 9.162228344046525, "frac_reward_zero_std": 0.17916667349636556, "grad_norm": 0.00665283203125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1532724636.0, "reward": 7.9718400001525875, "reward_std": 0.3321808859705925, "rewards/combined_reward/mean": 7.971839666366577, "rewards/combined_reward/std": 2.277187979221344, "step": 14970 }, { "clip_ratio/high_max": 0.0032007670709087205, "clip_ratio/high_mean": 0.001685526595247211, "clip_ratio/low_mean": 0.0009861419874747904, "clip_ratio/low_min": 0.00034977836985490286, "clip_ratio/region_mean": 0.0026716685524055114, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 183.72552642822265, "completions/mean_terminated_length": 183.72552642822265, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03215293204411864, "epoch": 9.180593816957453, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1535730302.0, "reward": 7.689777660369873, "reward_std": 0.33459707349538803, "rewards/combined_reward/mean": 7.689777040481568, "rewards/combined_reward/std": 2.7202370405197143, "step": 15000 }, { "clip_ratio/high_max": 0.0023965767560487922, "clip_ratio/high_mean": 0.0012966319414166112, "clip_ratio/low_mean": 0.0007192638791214752, "clip_ratio/low_min": 9.590360641595908e-05, "clip_ratio/region_mean": 0.002015895809745416, "completions/clipped_ratio": 0.0, "completions/max_length": 574.3, "completions/max_terminated_length": 574.3, "completions/mean_length": 197.75912017822264, "completions/mean_terminated_length": 197.75912017822264, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.02730826921761036, "epoch": 9.19895928986838, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1538789857.0, "reward": 8.091405296325684, "reward_std": 0.30578589290380476, "rewards/combined_reward/mean": 8.091405200958253, "rewards/combined_reward/std": 2.4103806972503663, "step": 15030 }, { "clip_ratio/high_max": 0.002765709865100992, "clip_ratio/high_mean": 0.0015555953798563375, "clip_ratio/low_mean": 0.0010258896077478615, "clip_ratio/low_min": 0.0002466386099210164, "clip_ratio/region_mean": 0.0025814849990031994, "completions/clipped_ratio": 0.0, "completions/max_length": 560.7, "completions/max_terminated_length": 560.7, "completions/mean_length": 205.73932800292968, "completions/mean_terminated_length": 205.73932800292968, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.029721138641859095, "epoch": 9.217324762779308, "frac_reward_zero_std": 0.14166666753590107, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1541880056.0, "reward": 8.058812618255615, "reward_std": 0.3187206253409386, "rewards/combined_reward/mean": 8.058812379837036, "rewards/combined_reward/std": 2.279843330383301, "step": 15060 }, { "clip_ratio/high_max": 0.0024902460354496725, "clip_ratio/high_mean": 0.0014315276629834747, "clip_ratio/low_mean": 0.0008900058920213876, "clip_ratio/low_min": 0.00021660948138257178, "clip_ratio/region_mean": 0.0023215335687079156, "completions/clipped_ratio": 0.0, "completions/max_length": 621.9, "completions/max_terminated_length": 621.9, "completions/mean_length": 196.63438262939454, "completions/mean_terminated_length": 196.63438262939454, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.029004097795890024, "epoch": 9.235690235690235, "frac_reward_zero_std": 0.14583333395421505, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1544935292.0, "reward": 8.030477905273438, "reward_std": 0.31239051148295405, "rewards/combined_reward/mean": 8.03047742843628, "rewards/combined_reward/std": 2.308894419670105, "step": 15090 }, { "clip_ratio/high_max": 0.002282070134970127, "clip_ratio/high_mean": 0.0013249860364643003, "clip_ratio/low_mean": 0.0009723433155159, "clip_ratio/low_min": 0.00027434601288405246, "clip_ratio/region_mean": 0.002297329337549551, "completions/clipped_ratio": 0.0, "completions/max_length": 688.2, "completions/max_terminated_length": 688.2, "completions/mean_length": 197.69088897705078, "completions/mean_terminated_length": 197.69088897705078, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.029838363950451214, "epoch": 9.254055708601163, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1547994585.0, "reward": 8.165716791152954, "reward_std": 0.31167530044913294, "rewards/combined_reward/mean": 8.165716218948365, "rewards/combined_reward/std": 2.081389236450195, "step": 15120 }, { "clip_ratio/high_max": 0.002480881041265093, "clip_ratio/high_mean": 0.0014323260097929354, "clip_ratio/low_mean": 0.0008985060222888327, "clip_ratio/low_min": 0.00023770467184173563, "clip_ratio/region_mean": 0.0023308320184393474, "completions/clipped_ratio": 0.0, "completions/max_length": 588.4, "completions/max_terminated_length": 588.4, "completions/mean_length": 206.0940185546875, "completions/mean_terminated_length": 206.0940185546875, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.029203310484687488, "epoch": 9.27242118151209, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1551086146.0, "reward": 8.274507188796997, "reward_std": 0.3314538553357124, "rewards/combined_reward/mean": 8.274506998062133, "rewards/combined_reward/std": 2.008057641983032, "step": 15150 }, { "clip_ratio/high_max": 0.002386805288551841, "clip_ratio/high_mean": 0.0013497872232013227, "clip_ratio/low_mean": 0.0008993563385350475, "clip_ratio/low_min": 0.00022741775028407575, "clip_ratio/region_mean": 0.002249143568527264, "completions/clipped_ratio": 0.0, "completions/max_length": 560.7, "completions/max_terminated_length": 560.7, "completions/mean_length": 196.99036712646483, "completions/mean_terminated_length": 196.99036712646483, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.027206052389616767, "epoch": 9.290786654423018, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1554142749.0, "reward": 8.156862354278564, "reward_std": 0.39378411024808885, "rewards/combined_reward/mean": 8.156861972808837, "rewards/combined_reward/std": 2.1453936457633973, "step": 15180 }, { "clip_ratio/high_max": 0.0022595549962716177, "clip_ratio/high_mean": 0.001273184357220695, "clip_ratio/low_mean": 0.0008502233679488806, "clip_ratio/low_min": 0.00014051336765987798, "clip_ratio/region_mean": 0.002123407720743368, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 194.6789123535156, "completions/mean_terminated_length": 194.6789123535156, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.026930076714294652, "epoch": 9.309152127333945, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1557190476.0, "reward": 8.231637334823608, "reward_std": 0.29393420368433, "rewards/combined_reward/mean": 8.231637287139893, "rewards/combined_reward/std": 1.9755455017089845, "step": 15210 }, { "clip_ratio/high_max": 0.0027434813668757367, "clip_ratio/high_mean": 0.0014345134933440326, "clip_ratio/low_mean": 0.0009342319472731712, "clip_ratio/low_min": 0.0001671598993804461, "clip_ratio/region_mean": 0.0023687454233974376, "completions/clipped_ratio": 0.0, "completions/max_length": 548.3, "completions/max_terminated_length": 548.3, "completions/mean_length": 204.11459045410157, "completions/mean_terminated_length": 204.11459045410157, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.031297785532660785, "epoch": 9.327517600244873, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1560274436.0, "reward": 7.947597169876099, "reward_std": 0.30586516335606573, "rewards/combined_reward/mean": 7.9475969791412355, "rewards/combined_reward/std": 2.5123722314834596, "step": 15240 }, { "clip_ratio/high_max": 0.002487238487810828, "clip_ratio/high_mean": 0.0014159451879095287, "clip_ratio/low_mean": 0.0008219061637646519, "clip_ratio/low_min": 0.00022759626105350132, "clip_ratio/region_mean": 0.002237851346338478, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 201.69349365234376, "completions/mean_terminated_length": 201.69349365234376, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02887465969348947, "epoch": 9.3458830731558, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1563349099.0, "reward": 8.000529384613037, "reward_std": 0.3100908428430557, "rewards/combined_reward/mean": 8.000528812408447, "rewards/combined_reward/std": 2.393703055381775, "step": 15270 }, { "clip_ratio/high_max": 0.0023961972246373383, "clip_ratio/high_mean": 0.0012734798447733434, "clip_ratio/low_mean": 0.0009327034105808707, "clip_ratio/low_min": 0.00024015466163594587, "clip_ratio/region_mean": 0.0022061832569306716, "completions/clipped_ratio": 0.0, "completions/max_length": 607.3, "completions/max_terminated_length": 607.3, "completions/mean_length": 200.16953735351564, "completions/mean_terminated_length": 200.16953735351564, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.028333080016697446, "epoch": 9.364248546066728, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1566417910.0, "reward": 8.1653742313385, "reward_std": 0.33950391262769697, "rewards/combined_reward/mean": 8.165373754501342, "rewards/combined_reward/std": 2.1145140647888185, "step": 15300 }, { "clip_ratio/high_max": 0.0023990788923886914, "clip_ratio/high_mean": 0.0014810292142404555, "clip_ratio/low_mean": 0.0009798142104045838, "clip_ratio/low_min": 0.0002748450632983198, "clip_ratio/region_mean": 0.0024608434313753, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 202.52787170410156, "completions/mean_terminated_length": 202.52787170410156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.028865340064900616, "epoch": 9.382614018977655, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0057373046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1569495777.0, "reward": 8.199349355697631, "reward_std": 0.29594150334596636, "rewards/combined_reward/mean": 8.19934902191162, "rewards/combined_reward/std": 2.038415789604187, "step": 15330 }, { "clip_ratio/high_max": 0.0029442382108148496, "clip_ratio/high_mean": 0.0016859343223283455, "clip_ratio/low_mean": 0.0009040922080506183, "clip_ratio/low_min": 0.00026767000041824456, "clip_ratio/region_mean": 0.002590026563363305, "completions/clipped_ratio": 0.0, "completions/max_length": 642.6, "completions/max_terminated_length": 642.6, "completions/mean_length": 201.4015670776367, "completions/mean_terminated_length": 201.4015670776367, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02968367306360354, "epoch": 9.400979491888583, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.00543212890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1572569319.0, "reward": 8.028643321990966, "reward_std": 0.3312353640794754, "rewards/combined_reward/mean": 8.028643131256104, "rewards/combined_reward/std": 2.2092148303985595, "step": 15360 }, { "clip_ratio/high_max": 0.002280790458219902, "clip_ratio/high_mean": 0.0012902710021686894, "clip_ratio/low_mean": 0.0009426318417657361, "clip_ratio/low_min": 0.0002448119778515926, "clip_ratio/region_mean": 0.0022329028472692397, "completions/clipped_ratio": 0.0, "completions/max_length": 671.1, "completions/max_terminated_length": 671.1, "completions/mean_length": 207.61875610351564, "completions/mean_terminated_length": 207.61875610351564, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.03012187658653905, "epoch": 9.41934496479951, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.007598876953125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1575666735.0, "reward": 7.87884373664856, "reward_std": 0.38667572736740113, "rewards/combined_reward/mean": 7.878843688964844, "rewards/combined_reward/std": 2.4677934408187867, "step": 15390 }, { "clip_ratio/high_max": 0.0020835160236553443, "clip_ratio/high_mean": 0.0011883607681132465, "clip_ratio/low_mean": 0.000773693565861322, "clip_ratio/low_min": 0.00017773668805602937, "clip_ratio/region_mean": 0.0019620543397953346, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 187.35208892822266, "completions/mean_terminated_length": 187.35208892822266, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.026065554888918995, "epoch": 9.437710437710438, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.01116943359375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1578686327.0, "reward": 8.330446863174439, "reward_std": 0.3031562313437462, "rewards/combined_reward/mean": 8.33044629096985, "rewards/combined_reward/std": 2.09073224067688, "step": 15420 }, { "clip_ratio/high_max": 0.0021669708633756577, "clip_ratio/high_mean": 0.0012076336524235861, "clip_ratio/low_mean": 0.0009948691773994749, "clip_ratio/low_min": 0.0002932262502630086, "clip_ratio/region_mean": 0.002202502820970646, "completions/clipped_ratio": 0.0, "completions/max_length": 583.6, "completions/max_terminated_length": 583.6, "completions/mean_length": 201.34401245117186, "completions/mean_terminated_length": 201.34401245117186, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.029254670611893137, "epoch": 9.456075910621365, "frac_reward_zero_std": 0.18750000633299352, "grad_norm": 0.0230712890625, "learning_rate": 5e-06, "loss": 0.0027, "num_tokens": 1581759648.0, "reward": 8.073462581634521, "reward_std": 0.3555474251508713, "rewards/combined_reward/mean": 8.073462057113648, "rewards/combined_reward/std": 2.374650037288666, "step": 15450 }, { "clip_ratio/high_max": 0.0025179587518020224, "clip_ratio/high_mean": 0.0014575888516750032, "clip_ratio/low_mean": 0.0008513361668519792, "clip_ratio/low_min": 0.0002013741543729945, "clip_ratio/region_mean": 0.002308925044295999, "completions/clipped_ratio": 0.0, "completions/max_length": 539.3, "completions/max_terminated_length": 539.3, "completions/mean_length": 203.39141235351562, "completions/mean_terminated_length": 203.39141235351562, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.027842297249784074, "epoch": 9.474441383532293, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.01226806640625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1584840831.0, "reward": 7.999658870697021, "reward_std": 0.32222963869571686, "rewards/combined_reward/mean": 7.99965877532959, "rewards/combined_reward/std": 2.323154020309448, "step": 15480 }, { "clip_ratio/high_max": 0.0022491587699429755, "clip_ratio/high_mean": 0.0012643169433431468, "clip_ratio/low_mean": 0.001040461499724188, "clip_ratio/low_min": 0.00040522414419683626, "clip_ratio/region_mean": 0.00230477843870176, "completions/clipped_ratio": 0.0, "completions/max_length": 554.6, "completions/max_terminated_length": 554.6, "completions/mean_length": 197.89922332763672, "completions/mean_terminated_length": 197.89922332763672, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02781168019088606, "epoch": 9.49280685644322, "frac_reward_zero_std": 0.19166667051613331, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 1587900924.0, "reward": 8.042923736572266, "reward_std": 0.3165672078728676, "rewards/combined_reward/mean": 8.04292345046997, "rewards/combined_reward/std": 2.3311051607131956, "step": 15510 }, { "clip_ratio/high_max": 0.001899847388995113, "clip_ratio/high_mean": 0.0011073985862215827, "clip_ratio/low_mean": 0.0009046713278318445, "clip_ratio/low_min": 0.00025102531241524655, "clip_ratio/region_mean": 0.0020120699162362145, "completions/clipped_ratio": 0.0, "completions/max_length": 523.4, "completions/max_terminated_length": 523.4, "completions/mean_length": 201.3463607788086, "completions/mean_terminated_length": 201.3463607788086, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.026077308428163328, "epoch": 9.511172329354148, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1590974254.0, "reward": 8.129623317718506, "reward_std": 0.30567166656255723, "rewards/combined_reward/mean": 8.129623222351075, "rewards/combined_reward/std": 2.3533190846443177, "step": 15540 }, { "clip_ratio/high_max": 0.0024065018233765537, "clip_ratio/high_mean": 0.0013768426942988299, "clip_ratio/low_mean": 0.0009808997909810083, "clip_ratio/low_min": 0.0002721053489949554, "clip_ratio/region_mean": 0.0023577424683026036, "completions/clipped_ratio": 0.0, "completions/max_length": 569.2, "completions/max_terminated_length": 569.2, "completions/mean_length": 207.89739837646485, "completions/mean_terminated_length": 207.89739837646485, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.029079265783851345, "epoch": 9.529537802265075, "frac_reward_zero_std": 0.19583334103226663, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1594072740.0, "reward": 8.01542296409607, "reward_std": 0.3365316055715084, "rewards/combined_reward/mean": 8.015423059463501, "rewards/combined_reward/std": 2.1930166125297545, "step": 15570 }, { "clip_ratio/high_max": 0.002619843782061556, "clip_ratio/high_mean": 0.001440330899034355, "clip_ratio/low_mean": 0.0010732871129827497, "clip_ratio/low_min": 0.0002830794013182943, "clip_ratio/region_mean": 0.002513618005711275, "completions/clipped_ratio": 0.0, "completions/max_length": 659.2, "completions/max_terminated_length": 659.2, "completions/mean_length": 199.6862030029297, "completions/mean_terminated_length": 199.6862030029297, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.03103625715399782, "epoch": 9.547903275176003, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.0098876953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1597139695.0, "reward": 8.060033798217773, "reward_std": 0.2849582314491272, "rewards/combined_reward/mean": 8.060033464431763, "rewards/combined_reward/std": 2.353414499759674, "step": 15600 }, { "clip_ratio/high_max": 0.0023317369627572286, "clip_ratio/high_mean": 0.00140836474262566, "clip_ratio/low_mean": 0.0009481299407828677, "clip_ratio/low_min": 0.00027064366392247997, "clip_ratio/region_mean": 0.0023564946881379, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 193.77031707763672, "completions/mean_terminated_length": 193.77031707763672, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.02958491452348729, "epoch": 9.56626874808693, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1600183933.0, "reward": 8.0665846824646, "reward_std": 0.32663758248090746, "rewards/combined_reward/mean": 8.06658411026001, "rewards/combined_reward/std": 2.292444384098053, "step": 15630 }, { "clip_ratio/high_max": 0.0022148235538528146, "clip_ratio/high_mean": 0.0012789464696349266, "clip_ratio/low_mean": 0.0009859020591344839, "clip_ratio/low_min": 0.0003605130977424172, "clip_ratio/region_mean": 0.002264848530224602, "completions/clipped_ratio": 0.0, "completions/max_length": 558.4, "completions/max_terminated_length": 558.4, "completions/mean_length": 186.1007873535156, "completions/mean_terminated_length": 186.1007873535156, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02888482214572529, "epoch": 9.584634220997858, "frac_reward_zero_std": 0.18333334103226662, "grad_norm": 0.02294921875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1603198720.0, "reward": 8.004857873916626, "reward_std": 0.25361480563879013, "rewards/combined_reward/mean": 8.004857873916626, "rewards/combined_reward/std": 2.418435847759247, "step": 15660 }, { "clip_ratio/high_max": 0.002964310406241566, "clip_ratio/high_mean": 0.0017063436192984228, "clip_ratio/low_mean": 0.00111750096778754, "clip_ratio/low_min": 0.0003281053657701705, "clip_ratio/region_mean": 0.0028238446150983996, "completions/clipped_ratio": 0.0, "completions/max_length": 592.2, "completions/max_terminated_length": 592.2, "completions/mean_length": 209.49193420410157, "completions/mean_terminated_length": 209.49193420410157, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03445508549921215, "epoch": 9.602999693908785, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1606303329.0, "reward": 7.86459174156189, "reward_std": 0.3432623207569122, "rewards/combined_reward/mean": 7.864591455459594, "rewards/combined_reward/std": 2.5179053425788878, "step": 15690 }, { "clip_ratio/high_max": 0.0024507431085415495, "clip_ratio/high_mean": 0.0013937496634753189, "clip_ratio/low_mean": 0.0010878706266036413, "clip_ratio/low_min": 0.000287143321717546, "clip_ratio/region_mean": 0.002481620296020992, "completions/clipped_ratio": 0.0, "completions/max_length": 550.2, "completions/max_terminated_length": 550.2, "completions/mean_length": 183.09037017822266, "completions/mean_terminated_length": 183.09037017822266, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03256815344405671, "epoch": 9.621365166819713, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1609306556.0, "reward": 8.065505504608154, "reward_std": 0.3626285523176193, "rewards/combined_reward/mean": 8.065505170822144, "rewards/combined_reward/std": 2.015766203403473, "step": 15720 }, { "clip_ratio/high_max": 0.002706084123929031, "clip_ratio/high_mean": 0.00162074305396042, "clip_ratio/low_mean": 0.0010288480420664807, "clip_ratio/low_min": 0.0003224230124033056, "clip_ratio/region_mean": 0.0026495911086385606, "completions/clipped_ratio": 0.0, "completions/max_length": 554.1, "completions/max_terminated_length": 554.1, "completions/mean_length": 203.73542022705078, "completions/mean_terminated_length": 203.73542022705078, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.0330269455909729, "epoch": 9.63973063973064, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.006683349609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1612389060.0, "reward": 8.175306272506713, "reward_std": 0.34584310799837115, "rewards/combined_reward/mean": 8.175305700302124, "rewards/combined_reward/std": 2.1540395855903625, "step": 15750 }, { "clip_ratio/high_max": 0.002666828960839969, "clip_ratio/high_mean": 0.0014330149327482407, "clip_ratio/low_mean": 0.0010148269324417925, "clip_ratio/low_min": 0.0002867766187894934, "clip_ratio/region_mean": 0.0024478418645837033, "completions/clipped_ratio": 0.0, "completions/max_length": 602.5, "completions/max_terminated_length": 602.5, "completions/mean_length": 205.13802490234374, "completions/mean_terminated_length": 205.13802490234374, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.03243626180725793, "epoch": 9.658096112641568, "frac_reward_zero_std": 0.16666667275130748, "grad_norm": 0.02294921875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1615476950.0, "reward": 7.7589616775512695, "reward_std": 0.4160821884870529, "rewards/combined_reward/mean": 7.75896143913269, "rewards/combined_reward/std": 2.4877548575401307, "step": 15780 }, { "clip_ratio/high_max": 0.002436037727359993, "clip_ratio/high_mean": 0.00130621413627523, "clip_ratio/low_mean": 0.0010940781165118096, "clip_ratio/low_min": 0.00027216490101030405, "clip_ratio/region_mean": 0.00240029225727388, "completions/clipped_ratio": 0.0, "completions/max_length": 593.5, "completions/max_terminated_length": 593.5, "completions/mean_length": 198.05912170410156, "completions/mean_terminated_length": 198.05912170410156, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.031854856045295796, "epoch": 9.676461585552495, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0034, "num_tokens": 1618537657.0, "reward": 8.128050422668457, "reward_std": 0.3684902131557465, "rewards/combined_reward/mean": 8.128050184249878, "rewards/combined_reward/std": 2.141422951221466, "step": 15810 }, { "clip_ratio/high_max": 0.002493177091916247, "clip_ratio/high_mean": 0.0014344150476972573, "clip_ratio/low_mean": 0.0011164969537882522, "clip_ratio/low_min": 0.00032278775494584504, "clip_ratio/region_mean": 0.002550912011550584, "completions/clipped_ratio": 0.0, "completions/max_length": 631.4, "completions/max_terminated_length": 631.4, "completions/mean_length": 188.15729675292968, "completions/mean_terminated_length": 188.15729675292968, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.031128897393743197, "epoch": 9.694827058463423, "frac_reward_zero_std": 0.2041666701436043, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1621560341.0, "reward": 8.055841779708862, "reward_std": 0.28954897820949554, "rewards/combined_reward/mean": 8.055841445922852, "rewards/combined_reward/std": 2.3205411434173584, "step": 15840 }, { "clip_ratio/high_max": 0.0022164042539467724, "clip_ratio/high_mean": 0.0012915353454445722, "clip_ratio/low_mean": 0.0008468981108308071, "clip_ratio/low_min": 0.00024373745133440632, "clip_ratio/region_mean": 0.002138433450212081, "completions/clipped_ratio": 0.0, "completions/max_length": 557.2, "completions/max_terminated_length": 557.2, "completions/mean_length": 202.84193267822266, "completions/mean_terminated_length": 202.84193267822266, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02924817972816527, "epoch": 9.71319253137435, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1624639414.0, "reward": 7.947677373886108, "reward_std": 0.3143570922315121, "rewards/combined_reward/mean": 7.947677373886108, "rewards/combined_reward/std": 2.3835056245326998, "step": 15870 }, { "clip_ratio/high_max": 0.002436047120136209, "clip_ratio/high_mean": 0.0014203934417309938, "clip_ratio/low_mean": 0.0010664272813301068, "clip_ratio/low_min": 0.00037201755282391483, "clip_ratio/region_mean": 0.002486820727547941, "completions/clipped_ratio": 0.0, "completions/max_length": 655.6, "completions/max_terminated_length": 655.6, "completions/mean_length": 201.21458740234374, "completions/mean_terminated_length": 201.21458740234374, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03180865189060569, "epoch": 9.731558004285278, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 1627712238.0, "reward": 8.025851964950562, "reward_std": 0.35462151020765303, "rewards/combined_reward/mean": 8.025851726531982, "rewards/combined_reward/std": 2.1726863265037535, "step": 15900 }, { "clip_ratio/high_max": 0.0030316341731425685, "clip_ratio/high_mean": 0.00173012320616787, "clip_ratio/low_mean": 0.0011025260947159646, "clip_ratio/low_min": 0.00033851897533168087, "clip_ratio/region_mean": 0.0028326493105851115, "completions/clipped_ratio": 0.0, "completions/max_length": 611.4, "completions/max_terminated_length": 611.4, "completions/mean_length": 209.68255767822265, "completions/mean_terminated_length": 209.68255767822265, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.034247903432697055, "epoch": 9.749923477196205, "frac_reward_zero_std": 0.16666667126119136, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1630817579.0, "reward": 7.613943386077881, "reward_std": 0.35145987272262574, "rewards/combined_reward/mean": 7.613942956924438, "rewards/combined_reward/std": 2.768878769874573, "step": 15930 }, { "clip_ratio/high_max": 0.0024464013118025226, "clip_ratio/high_mean": 0.0014678692189287782, "clip_ratio/low_mean": 0.0008247535984992283, "clip_ratio/low_min": 0.0002141646298696287, "clip_ratio/region_mean": 0.0022926228091819214, "completions/clipped_ratio": 0.0, "completions/max_length": 612.2, "completions/max_terminated_length": 612.2, "completions/mean_length": 187.03646240234374, "completions/mean_terminated_length": 187.03646240234374, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.029631104118501148, "epoch": 9.768288950107133, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.00726318359375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1633835959.0, "reward": 8.105064344406127, "reward_std": 0.28825646787881853, "rewards/combined_reward/mean": 8.105064105987548, "rewards/combined_reward/std": 2.320715272426605, "step": 15960 }, { "clip_ratio/high_max": 0.0024106973520247265, "clip_ratio/high_mean": 0.0013703031319892034, "clip_ratio/low_mean": 0.0009612384059437318, "clip_ratio/low_min": 0.00027139823578181677, "clip_ratio/region_mean": 0.0023315415334460945, "completions/clipped_ratio": 0.0, "completions/max_length": 557.3, "completions/max_terminated_length": 557.3, "completions/mean_length": 210.24687957763672, "completions/mean_terminated_length": 210.24687957763672, "completions/min_length": 48.3, "completions/min_terminated_length": 48.3, "entropy": 0.029836417486270268, "epoch": 9.78665442301806, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 1636943467.0, "reward": 8.132658195495605, "reward_std": 0.38014519661664964, "rewards/combined_reward/mean": 8.132658004760742, "rewards/combined_reward/std": 2.0085845947265626, "step": 15990 }, { "clip_ratio/high_max": 0.0024193981970408156, "clip_ratio/high_mean": 0.001399264277764208, "clip_ratio/low_mean": 0.0011570879408585217, "clip_ratio/low_min": 0.00040626664801190294, "clip_ratio/region_mean": 0.00255635225524505, "completions/clipped_ratio": 0.0, "completions/max_length": 644.0, "completions/max_terminated_length": 644.0, "completions/mean_length": 195.4869857788086, "completions/mean_terminated_length": 195.4869857788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03132679252885282, "epoch": 9.805019895928988, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1639994297.0, "reward": 8.022938632965088, "reward_std": 0.3540364310145378, "rewards/combined_reward/mean": 8.022938585281372, "rewards/combined_reward/std": 2.0767744362354277, "step": 16020 }, { "clip_ratio/high_max": 0.0027077021816997635, "clip_ratio/high_mean": 0.0015323863013084822, "clip_ratio/low_mean": 0.0009234705839238207, "clip_ratio/low_min": 0.00021827632881468161, "clip_ratio/region_mean": 0.0024558568741971007, "completions/clipped_ratio": 0.0, "completions/max_length": 568.5, "completions/max_terminated_length": 568.5, "completions/mean_length": 213.33307647705078, "completions/mean_terminated_length": 213.33307647705078, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02992872764977316, "epoch": 9.823385368839915, "frac_reward_zero_std": 0.15000000186264514, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1643113656.0, "reward": 7.799420118331909, "reward_std": 0.4105433076620102, "rewards/combined_reward/mean": 7.7994202136993405, "rewards/combined_reward/std": 2.5533790230751037, "step": 16050 }, { "clip_ratio/high_max": 0.001972080383954259, "clip_ratio/high_mean": 0.0010374293335189578, "clip_ratio/low_mean": 0.0010410875288774453, "clip_ratio/low_min": 0.0003391373727936298, "clip_ratio/region_mean": 0.0020785168434182804, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 195.7182357788086, "completions/mean_terminated_length": 195.7182357788086, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.026117548517261942, "epoch": 9.841750841750843, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.0140380859375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1646165374.0, "reward": 7.893317031860351, "reward_std": 0.30487425774335863, "rewards/combined_reward/mean": 7.893317031860351, "rewards/combined_reward/std": 2.34610995054245, "step": 16080 }, { "clip_ratio/high_max": 0.002280764556538391, "clip_ratio/high_mean": 0.0012806943347338044, "clip_ratio/low_mean": 0.0009748727909027366, "clip_ratio/low_min": 0.00026562641481480873, "clip_ratio/region_mean": 0.0022555671283043923, "completions/clipped_ratio": 0.0, "completions/max_length": 665.3, "completions/max_terminated_length": 665.3, "completions/mean_length": 205.46979827880858, "completions/mean_terminated_length": 205.46979827880858, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02733784761900703, "epoch": 9.86011631466177, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1649254538.0, "reward": 7.925510454177856, "reward_std": 0.31943133994936945, "rewards/combined_reward/mean": 7.9255101680755615, "rewards/combined_reward/std": 2.3638421297073364, "step": 16110 }, { "clip_ratio/high_max": 0.0026948774752478736, "clip_ratio/high_mean": 0.0014759387115191203, "clip_ratio/low_mean": 0.0010823560445714976, "clip_ratio/low_min": 0.00031831211720903714, "clip_ratio/region_mean": 0.0025582947419025004, "completions/clipped_ratio": 0.0, "completions/max_length": 586.3, "completions/max_terminated_length": 586.3, "completions/mean_length": 202.6627655029297, "completions/mean_terminated_length": 202.6627655029297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03072200873090575, "epoch": 9.878481787572696, "frac_reward_zero_std": 0.15000000298023225, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1652332923.0, "reward": 8.001695775985718, "reward_std": 0.3557304561138153, "rewards/combined_reward/mean": 8.001695489883422, "rewards/combined_reward/std": 2.445793867111206, "step": 16140 }, { "clip_ratio/high_max": 0.002586278675880749, "clip_ratio/high_mean": 0.001522355465082607, "clip_ratio/low_mean": 0.0009569039753235605, "clip_ratio/low_min": 0.00028492931160144507, "clip_ratio/region_mean": 0.0024792594311293216, "completions/clipped_ratio": 0.0, "completions/max_length": 581.1, "completions/max_terminated_length": 581.1, "completions/mean_length": 211.05573425292968, "completions/mean_terminated_length": 211.05573425292968, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 0.02991026609670371, "epoch": 9.896847260483623, "frac_reward_zero_std": 0.1291666690260172, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1655443537.0, "reward": 7.985467910766602, "reward_std": 0.446404767036438, "rewards/combined_reward/mean": 7.985467481613159, "rewards/combined_reward/std": 2.3043187975883486, "step": 16170 }, { "clip_ratio/high_max": 0.002204239111839949, "clip_ratio/high_mean": 0.0012984975595221233, "clip_ratio/low_mean": 0.0009994197670797197, "clip_ratio/low_min": 0.0003031273108111539, "clip_ratio/region_mean": 0.002297917314475247, "completions/clipped_ratio": 0.0, "completions/max_length": 615.8, "completions/max_terminated_length": 615.8, "completions/mean_length": 215.3502670288086, "completions/mean_terminated_length": 215.3502670288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029436961219956478, "epoch": 9.91521273339455, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1658570642.0, "reward": 7.955909204483032, "reward_std": 0.3681539326906204, "rewards/combined_reward/mean": 7.9559091091156, "rewards/combined_reward/std": 2.3078439116477965, "step": 16200 }, { "clip_ratio/high_max": 0.002425825582273925, "clip_ratio/high_mean": 0.00134599918492313, "clip_ratio/low_mean": 0.001000842605693227, "clip_ratio/low_min": 0.00023334224558008526, "clip_ratio/region_mean": 0.002346841797407251, "completions/clipped_ratio": 0.0, "completions/max_length": 605.5, "completions/max_terminated_length": 605.5, "completions/mean_length": 204.72266235351563, "completions/mean_terminated_length": 204.72266235351563, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.02819487953869005, "epoch": 9.933578206305478, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1661656937.0, "reward": 8.169746685028077, "reward_std": 0.379040589928627, "rewards/combined_reward/mean": 8.169746351242065, "rewards/combined_reward/std": 2.297114408016205, "step": 16230 }, { "clip_ratio/high_max": 0.002706336065117891, "clip_ratio/high_mean": 0.0016466736529158273, "clip_ratio/low_mean": 0.001061883809961728, "clip_ratio/low_min": 0.0003644349211147831, "clip_ratio/region_mean": 0.002708557472215034, "completions/clipped_ratio": 0.0, "completions/max_length": 615.8, "completions/max_terminated_length": 615.8, "completions/mean_length": 209.3205780029297, "completions/mean_terminated_length": 209.3205780029297, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.03152994709089398, "epoch": 9.951943679216406, "frac_reward_zero_std": 0.17083333469927311, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1664760888.0, "reward": 7.818707895278931, "reward_std": 0.4083564430475235, "rewards/combined_reward/mean": 7.818707752227783, "rewards/combined_reward/std": 2.4586417436599732, "step": 16260 }, { "clip_ratio/high_max": 0.0024971059120919865, "clip_ratio/high_mean": 0.0015073433467478025, "clip_ratio/low_mean": 0.0010002420751940614, "clip_ratio/low_min": 0.00035271583813785884, "clip_ratio/region_mean": 0.002507585418425151, "completions/clipped_ratio": 0.0, "completions/max_length": 529.1, "completions/max_terminated_length": 529.1, "completions/mean_length": 204.03802337646485, "completions/mean_terminated_length": 204.03802337646485, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.032684538730730614, "epoch": 9.970309152127333, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1667844554.0, "reward": 7.955766916275024, "reward_std": 0.3417156293988228, "rewards/combined_reward/mean": 7.955766773223877, "rewards/combined_reward/std": 2.3654021263122558, "step": 16290 }, { "clip_ratio/high_max": 0.002481117185379844, "clip_ratio/high_mean": 0.0014067898255840797, "clip_ratio/low_mean": 0.0008921874934458174, "clip_ratio/low_min": 0.0002237613439016665, "clip_ratio/region_mean": 0.0022989773271547163, "completions/clipped_ratio": 0.0, "completions/max_length": 530.8, "completions/max_terminated_length": 530.8, "completions/mean_length": 197.6565155029297, "completions/mean_terminated_length": 197.6565155029297, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.026853341360886893, "epoch": 9.98867462503826, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.0302734375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1670903715.0, "reward": 8.090933799743652, "reward_std": 0.3753828063607216, "rewards/combined_reward/mean": 8.0909339427948, "rewards/combined_reward/std": 2.4000354409217834, "step": 16320 }, { "clip_ratio/high_max": 0.002678110657390049, "clip_ratio/high_mean": 0.001479428832496169, "clip_ratio/low_mean": 0.0010677369036938664, "clip_ratio/low_min": 0.0002966624340811997, "clip_ratio/region_mean": 0.00254716570005689, "completions/clipped_ratio": 0.0, "completions/max_length": 533.2222222222222, "completions/max_terminated_length": 533.2222222222222, "completions/mean_length": 195.45023600260416, "completions/mean_terminated_length": 195.45023600260416, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.03432565351348307, "epoch": 10.006734006734007, "frac_reward_zero_std": 0.13425926326049697, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1673649335.0, "reward": 8.17202705807156, "reward_std": 0.4091225233342912, "rewards/combined_reward/mean": 8.172026422288683, "rewards/combined_reward/std": 2.1150901317596436, "step": 16350 }, { "clip_ratio/high_max": 0.001972304220301642, "clip_ratio/high_mean": 0.0010951141643090524, "clip_ratio/low_mean": 0.0007241395064435589, "clip_ratio/low_min": 0.00018814492141245865, "clip_ratio/region_mean": 0.0018192536765127443, "completions/clipped_ratio": 0.0, "completions/max_length": 706.4, "completions/max_terminated_length": 706.4, "completions/mean_length": 196.29662017822267, "completions/mean_terminated_length": 196.29662017822267, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.030982295190915464, "epoch": 10.025099479644934, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.0155029296875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1676703274.0, "reward": 7.94876070022583, "reward_std": 0.3101317957043648, "rewards/combined_reward/mean": 7.948760461807251, "rewards/combined_reward/std": 2.388523578643799, "step": 16380 }, { "clip_ratio/high_max": 0.0017463404190493747, "clip_ratio/high_mean": 0.0010127773508429526, "clip_ratio/low_mean": 0.0006371894909534603, "clip_ratio/low_min": 0.0001568864201544784, "clip_ratio/region_mean": 0.001649966833065264, "completions/clipped_ratio": 0.0, "completions/max_length": 590.75, "completions/max_terminated_length": 590.75, "completions/mean_length": 190.42188262939453, "completions/mean_terminated_length": 190.42188262939453, "completions/min_length": 38.25, "completions/min_terminated_length": 38.25, "entropy": 0.032777313888072965, "epoch": 10.042852770125497, "frac_reward_zero_std": 0.2500000111758709, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1680082308.0, "reward": 8.138707399368286, "reward_std": 0.2556082382798195, "rewards/combined_reward/mean": 8.138706922531128, "rewards/combined_reward/std": 2.289253145456314, "step": 16410 }, { "clip_ratio/high_max": 0.002312371319082255, "clip_ratio/high_mean": 0.0013063236927943458, "clip_ratio/low_mean": 0.0007981719577704401, "clip_ratio/low_min": 0.00023320018784337055, "clip_ratio/region_mean": 0.0021044956617212542, "completions/clipped_ratio": 0.0, "completions/max_length": 528.5, "completions/max_terminated_length": 528.5, "completions/mean_length": 185.21927490234376, "completions/mean_terminated_length": 185.21927490234376, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.030653623476003606, "epoch": 10.061218243036425, "frac_reward_zero_std": 0.22500000670552253, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1683093710.0, "reward": 7.992503261566162, "reward_std": 0.3393340528011322, "rewards/combined_reward/mean": 7.99250283241272, "rewards/combined_reward/std": 2.372429835796356, "step": 16440 }, { "clip_ratio/high_max": 0.0015102049403746302, "clip_ratio/high_mean": 0.0009147207587375305, "clip_ratio/low_mean": 0.0008703565718557609, "clip_ratio/low_min": 0.00032638823225473367, "clip_ratio/region_mean": 0.0017850773467216641, "completions/clipped_ratio": 0.0, "completions/max_length": 610.2, "completions/max_terminated_length": 610.2, "completions/mean_length": 201.30651397705077, "completions/mean_terminated_length": 201.30651397705077, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.028054037146891156, "epoch": 10.079583715947352, "frac_reward_zero_std": 0.13750000186264516, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1686166887.0, "reward": 7.902202367782593, "reward_std": 0.3486507087945938, "rewards/combined_reward/mean": 7.902202081680298, "rewards/combined_reward/std": 2.5936043739318846, "step": 16470 }, { "clip_ratio/high_max": 0.0017776228123693727, "clip_ratio/high_mean": 0.0009180772006705714, "clip_ratio/low_mean": 0.000760664385961718, "clip_ratio/low_min": 0.0002099689047706003, "clip_ratio/region_mean": 0.0016787415928168532, "completions/clipped_ratio": 0.0, "completions/max_length": 551.6, "completions/max_terminated_length": 551.6, "completions/mean_length": 197.08281555175782, "completions/mean_terminated_length": 197.08281555175782, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.027520990208722652, "epoch": 10.09794918885828, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.01507568359375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1689223845.0, "reward": 8.085558319091797, "reward_std": 0.2583959072828293, "rewards/combined_reward/mean": 8.08555793762207, "rewards/combined_reward/std": 2.020742416381836, "step": 16500 }, { "clip_ratio/high_max": 0.0018401700266016026, "clip_ratio/high_mean": 0.0009973366890335456, "clip_ratio/low_mean": 0.0008342612657240049, "clip_ratio/low_min": 0.00027485071817257755, "clip_ratio/region_mean": 0.0018315979830125192, "completions/clipped_ratio": 0.0, "completions/max_length": 749.9, "completions/max_terminated_length": 749.9, "completions/mean_length": 203.3921920776367, "completions/mean_terminated_length": 203.3921920776367, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.03186939219012856, "epoch": 10.116314661769207, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1692305031.0, "reward": 7.727038717269897, "reward_std": 0.33339186012744904, "rewards/combined_reward/mean": 7.727038431167602, "rewards/combined_reward/std": 2.625330424308777, "step": 16530 }, { "clip_ratio/high_max": 0.0018991951607555773, "clip_ratio/high_mean": 0.001108498689791304, "clip_ratio/low_mean": 0.0007271577616847936, "clip_ratio/low_min": 0.00024525601305261564, "clip_ratio/region_mean": 0.0018356564692415607, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 200.84739990234374, "completions/mean_terminated_length": 200.84739990234374, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03153010966877143, "epoch": 10.134680134680135, "frac_reward_zero_std": 0.17916667349636556, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1695376445.0, "reward": 8.042805337905884, "reward_std": 0.37204529494047167, "rewards/combined_reward/mean": 8.042805099487305, "rewards/combined_reward/std": 2.3165699362754824, "step": 16560 }, { "clip_ratio/high_max": 0.0018838232611111986, "clip_ratio/high_mean": 0.0010953730755621411, "clip_ratio/low_mean": 0.0007306658898111588, "clip_ratio/low_min": 0.00013247995035878073, "clip_ratio/region_mean": 0.0018260389852609174, "completions/clipped_ratio": 0.0, "completions/max_length": 522.8, "completions/max_terminated_length": 522.8, "completions/mean_length": 195.99115142822265, "completions/mean_terminated_length": 195.99115142822265, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.028961298195645212, "epoch": 10.153045607591062, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1698429211.0, "reward": 8.187817049026489, "reward_std": 0.29668440073728564, "rewards/combined_reward/mean": 8.187816762924195, "rewards/combined_reward/std": 2.0218950629234316, "step": 16590 }, { "clip_ratio/high_max": 0.0020321239081871074, "clip_ratio/high_mean": 0.0011723364583303918, "clip_ratio/low_mean": 0.0007873593822296243, "clip_ratio/low_min": 0.0002707371037104167, "clip_ratio/region_mean": 0.0019596958489273677, "completions/clipped_ratio": 0.0, "completions/max_length": 582.6, "completions/max_terminated_length": 582.6, "completions/mean_length": 195.33307952880858, "completions/mean_terminated_length": 195.33307952880858, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03142217467539012, "epoch": 10.17141108050199, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1701479450.0, "reward": 8.035321950912476, "reward_std": 0.38004328310489655, "rewards/combined_reward/mean": 8.035321712493896, "rewards/combined_reward/std": 2.4515037536621094, "step": 16620 }, { "clip_ratio/high_max": 0.0019698466334375554, "clip_ratio/high_mean": 0.0011135738727413506, "clip_ratio/low_mean": 0.0007148396249249345, "clip_ratio/low_min": 0.0001353664284882446, "clip_ratio/region_mean": 0.001828413497666285, "completions/clipped_ratio": 0.0, "completions/max_length": 578.7, "completions/max_terminated_length": 578.7, "completions/mean_length": 196.1851608276367, "completions/mean_terminated_length": 196.1851608276367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.0280965173849836, "epoch": 10.189776553412917, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1704532961.0, "reward": 8.078085327148438, "reward_std": 0.3370314806699753, "rewards/combined_reward/mean": 8.078085279464721, "rewards/combined_reward/std": 2.209067475795746, "step": 16650 }, { "clip_ratio/high_max": 0.0019895612810311527, "clip_ratio/high_mean": 0.0011298099445411936, "clip_ratio/low_mean": 0.0006922120884458612, "clip_ratio/low_min": 0.00019686505887269354, "clip_ratio/region_mean": 0.001822022029470342, "completions/clipped_ratio": 0.0, "completions/max_length": 555.8, "completions/max_terminated_length": 555.8, "completions/mean_length": 196.8377670288086, "completions/mean_terminated_length": 196.8377670288086, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027467424453546604, "epoch": 10.208142026323845, "frac_reward_zero_std": 0.24166667461395264, "grad_norm": 0.0257568359375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1707588978.0, "reward": 7.865289878845215, "reward_std": 0.31764726489782336, "rewards/combined_reward/mean": 7.865289545059204, "rewards/combined_reward/std": 2.5752118587493897, "step": 16680 }, { "clip_ratio/high_max": 0.00183940498221394, "clip_ratio/high_mean": 0.0011050474450409334, "clip_ratio/low_mean": 0.0008077257516561076, "clip_ratio/low_min": 0.00022550794140746195, "clip_ratio/region_mean": 0.0019127732220416268, "completions/clipped_ratio": 0.0, "completions/max_length": 543.9, "completions/max_terminated_length": 543.9, "completions/mean_length": 200.03594207763672, "completions/mean_terminated_length": 200.03594207763672, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.029510750761255623, "epoch": 10.226507499234772, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1710657276.0, "reward": 8.04518256187439, "reward_std": 0.3815748929977417, "rewards/combined_reward/mean": 8.045182228088379, "rewards/combined_reward/std": 2.202501356601715, "step": 16710 }, { "clip_ratio/high_max": 0.0017448373335355427, "clip_ratio/high_mean": 0.000976745773308115, "clip_ratio/low_mean": 0.0006649688376152578, "clip_ratio/low_min": 0.0001602393967914395, "clip_ratio/region_mean": 0.001641714612924261, "completions/clipped_ratio": 0.0, "completions/max_length": 571.9, "completions/max_terminated_length": 571.9, "completions/mean_length": 188.76693267822264, "completions/mean_terminated_length": 188.76693267822264, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.027484288532286884, "epoch": 10.2448729721457, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.0034942626953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1713682301.0, "reward": 8.161470746994018, "reward_std": 0.24493734985589982, "rewards/combined_reward/mean": 8.161470317840577, "rewards/combined_reward/std": 2.10980761051178, "step": 16740 }, { "clip_ratio/high_max": 0.0021647356528167925, "clip_ratio/high_mean": 0.001185017423267709, "clip_ratio/low_mean": 0.0007761497054161737, "clip_ratio/low_min": 0.000224002992278353, "clip_ratio/region_mean": 0.001961167117891212, "completions/clipped_ratio": 0.0, "completions/max_length": 543.2, "completions/max_terminated_length": 543.2, "completions/mean_length": 195.6080780029297, "completions/mean_terminated_length": 195.6080780029297, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.031610856569993, "epoch": 10.263238445056627, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.026123046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1716733596.0, "reward": 8.071628999710082, "reward_std": 0.4058314934372902, "rewards/combined_reward/mean": 8.07162880897522, "rewards/combined_reward/std": 2.2358471274375917, "step": 16770 }, { "clip_ratio/high_max": 0.0018430022328781584, "clip_ratio/high_mean": 0.001076438038580818, "clip_ratio/low_mean": 0.0008862001039233291, "clip_ratio/low_min": 0.00020695709439072136, "clip_ratio/region_mean": 0.0019626381414127535, "completions/clipped_ratio": 0.0, "completions/max_length": 603.4, "completions/max_terminated_length": 603.4, "completions/mean_length": 207.33802642822266, "completions/mean_terminated_length": 207.33802642822266, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02904058378189802, "epoch": 10.281603917967555, "frac_reward_zero_std": 0.16250000670552253, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 1719829934.0, "reward": 8.099327993392944, "reward_std": 0.3537015825510025, "rewards/combined_reward/mean": 8.099327564239502, "rewards/combined_reward/std": 2.2412218689918517, "step": 16800 }, { "clip_ratio/high_max": 0.0018407887890740918, "clip_ratio/high_mean": 0.0011212862426570305, "clip_ratio/low_mean": 0.0007371670816307111, "clip_ratio/low_min": 0.00020093324152791562, "clip_ratio/region_mean": 0.001858453308523167, "completions/clipped_ratio": 0.0, "completions/max_length": 550.5, "completions/max_terminated_length": 550.5, "completions/mean_length": 206.86302795410157, "completions/mean_terminated_length": 206.86302795410157, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.031444299841920535, "epoch": 10.299969390878482, "frac_reward_zero_std": 0.15833333991467952, "grad_norm": 0.00640869140625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1722924448.0, "reward": 7.9911261081695555, "reward_std": 0.2820972174406052, "rewards/combined_reward/mean": 7.99112606048584, "rewards/combined_reward/std": 2.1761693358421326, "step": 16830 }, { "clip_ratio/high_max": 0.0018469462971552274, "clip_ratio/high_mean": 0.00102871805477965, "clip_ratio/low_mean": 0.0007519776977763589, "clip_ratio/low_min": 0.00025898532330757005, "clip_ratio/region_mean": 0.0017806957558301898, "completions/clipped_ratio": 0.0, "completions/max_length": 631.8, "completions/max_terminated_length": 631.8, "completions/mean_length": 206.2359420776367, "completions/mean_terminated_length": 206.2359420776367, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.026584306359291078, "epoch": 10.31833486378941, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1726016554.0, "reward": 7.879658651351929, "reward_std": 0.29949782192707064, "rewards/combined_reward/mean": 7.879658317565918, "rewards/combined_reward/std": 2.3764671087265015, "step": 16860 }, { "clip_ratio/high_max": 0.0020698965866661942, "clip_ratio/high_mean": 0.0010945494592306203, "clip_ratio/low_mean": 0.0007793999811838148, "clip_ratio/low_min": 0.00015842729141392434, "clip_ratio/region_mean": 0.0018739494339873393, "completions/clipped_ratio": 0.0, "completions/max_length": 556.6, "completions/max_terminated_length": 556.6, "completions/mean_length": 198.46953582763672, "completions/mean_terminated_length": 198.46953582763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03128732222442825, "epoch": 10.336700336700337, "frac_reward_zero_std": 0.18750000558793545, "grad_norm": 0.048828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1729078837.0, "reward": 8.098320245742798, "reward_std": 0.34012997299432757, "rewards/combined_reward/mean": 8.098320055007935, "rewards/combined_reward/std": 2.0881091594696044, "step": 16890 }, { "clip_ratio/high_max": 0.002213156537375956, "clip_ratio/high_mean": 0.0012288068800747473, "clip_ratio/low_mean": 0.0008998076787975151, "clip_ratio/low_min": 0.0002786517596784203, "clip_ratio/region_mean": 0.002128614563359103, "completions/clipped_ratio": 0.0, "completions/max_length": 564.1, "completions/max_terminated_length": 564.1, "completions/mean_length": 204.4205780029297, "completions/mean_terminated_length": 204.4205780029297, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.03297534952871502, "epoch": 10.355065809611265, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1732163972.0, "reward": 8.04945774078369, "reward_std": 0.3076328054070473, "rewards/combined_reward/mean": 8.04945774078369, "rewards/combined_reward/std": 2.387988018989563, "step": 16920 }, { "clip_ratio/high_max": 0.002361245012434665, "clip_ratio/high_mean": 0.0014115437040648734, "clip_ratio/low_mean": 0.000880902321296162, "clip_ratio/low_min": 0.00026201702688316195, "clip_ratio/region_mean": 0.002292446017963812, "completions/clipped_ratio": 0.0, "completions/max_length": 553.6, "completions/max_terminated_length": 553.6, "completions/mean_length": 202.12214050292968, "completions/mean_terminated_length": 202.12214050292968, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.032355784655859074, "epoch": 10.373431282522192, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.005889892578125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1735240281.0, "reward": 8.027611255645752, "reward_std": 0.38216800391674044, "rewards/combined_reward/mean": 8.027610731124877, "rewards/combined_reward/std": 2.336776411533356, "step": 16950 }, { "clip_ratio/high_max": 0.0019279194263314519, "clip_ratio/high_mean": 0.0010679089123489878, "clip_ratio/low_mean": 0.00079131040389863, "clip_ratio/low_min": 0.00021813468022931679, "clip_ratio/region_mean": 0.0018592193082440645, "completions/clipped_ratio": 0.0, "completions/max_length": 693.1, "completions/max_terminated_length": 693.1, "completions/mean_length": 210.28985137939452, "completions/mean_terminated_length": 210.28985137939452, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02684107922638456, "epoch": 10.39179675543312, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1738347954.0, "reward": 8.010380220413207, "reward_std": 0.40038844645023347, "rewards/combined_reward/mean": 8.010379981994628, "rewards/combined_reward/std": 2.2711226105690003, "step": 16980 }, { "clip_ratio/high_max": 0.0021489285793601693, "clip_ratio/high_mean": 0.0011915210080284547, "clip_ratio/low_mean": 0.000776899472354368, "clip_ratio/low_min": 0.00027300318664250274, "clip_ratio/region_mean": 0.0019684204817167483, "completions/clipped_ratio": 0.0, "completions/max_length": 604.8, "completions/max_terminated_length": 604.8, "completions/mean_length": 197.8658920288086, "completions/mean_terminated_length": 197.8658920288086, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.030480919933567444, "epoch": 10.410162228344047, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1741407919.0, "reward": 8.278060102462769, "reward_std": 0.28219424188137054, "rewards/combined_reward/mean": 8.278059768676759, "rewards/combined_reward/std": 2.063336443901062, "step": 17010 }, { "clip_ratio/high_max": 0.001802560597328314, "clip_ratio/high_mean": 0.0011504792870255186, "clip_ratio/low_mean": 0.00099177955792887, "clip_ratio/low_min": 0.0003186050074873492, "clip_ratio/region_mean": 0.0021422588295536118, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 202.12787017822265, "completions/mean_terminated_length": 202.12787017822265, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.03533020803394417, "epoch": 10.428527701254975, "frac_reward_zero_std": 0.15833333544433117, "grad_norm": 0.036865234375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1744484250.0, "reward": 7.898676586151123, "reward_std": 0.42905998378992083, "rewards/combined_reward/mean": 7.898676586151123, "rewards/combined_reward/std": 2.444567310810089, "step": 17040 }, { "clip_ratio/high_max": 0.0019230897167290095, "clip_ratio/high_mean": 0.001138528994609563, "clip_ratio/low_mean": 0.0009079522156146898, "clip_ratio/low_min": 0.0003249286853436691, "clip_ratio/region_mean": 0.0020464811813629543, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 190.14037017822267, "completions/mean_terminated_length": 190.14037017822267, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029730791753778855, "epoch": 10.4468931741659, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1747514549.0, "reward": 7.906714344024659, "reward_std": 0.31522239595651624, "rewards/combined_reward/mean": 7.9067141056060795, "rewards/combined_reward/std": 2.3242304742336275, "step": 17070 }, { "clip_ratio/high_max": 0.0018757345746659364, "clip_ratio/high_mean": 0.00098358272856179, "clip_ratio/low_mean": 0.0007494067980587715, "clip_ratio/low_min": 0.0001313912670108645, "clip_ratio/region_mean": 0.0017329895238314445, "completions/clipped_ratio": 0.0, "completions/max_length": 574.2, "completions/max_terminated_length": 574.2, "completions/mean_length": 195.75938110351564, "completions/mean_terminated_length": 195.75938110351564, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.027522098800788323, "epoch": 10.465258647076828, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1750566425.0, "reward": 7.9741967678070065, "reward_std": 0.38082679361104965, "rewards/combined_reward/mean": 7.974196720123291, "rewards/combined_reward/std": 2.2879075407981873, "step": 17100 }, { "clip_ratio/high_max": 0.0017067231203933867, "clip_ratio/high_mean": 0.0009565399983330281, "clip_ratio/low_mean": 0.0008112080130861916, "clip_ratio/low_min": 0.00021157215887797066, "clip_ratio/region_mean": 0.001767748010267193, "completions/clipped_ratio": 0.0, "completions/max_length": 611.5, "completions/max_terminated_length": 611.5, "completions/mean_length": 201.74375762939454, "completions/mean_terminated_length": 201.74375762939454, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.027663994766771793, "epoch": 10.483624119987756, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.035400390625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1753641281.0, "reward": 8.195217657089234, "reward_std": 0.2648642875254154, "rewards/combined_reward/mean": 8.19521722793579, "rewards/combined_reward/std": 2.289850616455078, "step": 17130 }, { "clip_ratio/high_max": 0.0022018857353638547, "clip_ratio/high_mean": 0.0011277030663525997, "clip_ratio/low_mean": 0.0008829489687438278, "clip_ratio/low_min": 0.00028467162604404924, "clip_ratio/region_mean": 0.0020106519989591715, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 199.02943420410156, "completions/mean_terminated_length": 199.02943420410156, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03137787951466938, "epoch": 10.501989592898683, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.02001953125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1756705714.0, "reward": 7.894499254226685, "reward_std": 0.3409413084387779, "rewards/combined_reward/mean": 7.894498872756958, "rewards/combined_reward/std": 2.385008668899536, "step": 17160 }, { "clip_ratio/high_max": 0.001878055784618482, "clip_ratio/high_mean": 0.0010721368785501302, "clip_ratio/low_mean": 0.0007728485972620547, "clip_ratio/low_min": 0.00021041070237212505, "clip_ratio/region_mean": 0.0018449854716891423, "completions/clipped_ratio": 0.0, "completions/max_length": 644.8, "completions/max_terminated_length": 644.8, "completions/mean_length": 195.82396240234374, "completions/mean_terminated_length": 195.82396240234374, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03010026483486096, "epoch": 10.52035506580961, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1759757838.0, "reward": 7.861202907562256, "reward_std": 0.33028253465890883, "rewards/combined_reward/mean": 7.861202716827393, "rewards/combined_reward/std": 2.575404417514801, "step": 17190 }, { "clip_ratio/high_max": 0.0019437582457612734, "clip_ratio/high_mean": 0.0011859037931571947, "clip_ratio/low_mean": 0.0007651972773601301, "clip_ratio/low_min": 0.00022292040011961944, "clip_ratio/region_mean": 0.0019511010653028885, "completions/clipped_ratio": 0.0, "completions/max_length": 583.2, "completions/max_terminated_length": 583.2, "completions/mean_length": 198.6862030029297, "completions/mean_terminated_length": 198.6862030029297, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.030437767660866182, "epoch": 10.538720538720538, "frac_reward_zero_std": 0.1500000014901161, "grad_norm": 0.022705078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1762820953.0, "reward": 8.027050638198853, "reward_std": 0.30103936940431597, "rewards/combined_reward/mean": 8.027050399780274, "rewards/combined_reward/std": 2.324760854244232, "step": 17220 }, { "clip_ratio/high_max": 0.0019070346374064685, "clip_ratio/high_mean": 0.0011039775219008638, "clip_ratio/low_mean": 0.0006153641580264473, "clip_ratio/low_min": 0.00011086590311606414, "clip_ratio/region_mean": 0.0017193416938728963, "completions/clipped_ratio": 0.0, "completions/max_length": 567.5, "completions/max_terminated_length": 567.5, "completions/mean_length": 197.41094360351562, "completions/mean_terminated_length": 197.41094360351562, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.025473867449909447, "epoch": 10.557086011631466, "frac_reward_zero_std": 0.20416667275130748, "grad_norm": 0.004974365234375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1765879171.0, "reward": 8.347170448303222, "reward_std": 0.2546895295381546, "rewards/combined_reward/mean": 8.347170305252074, "rewards/combined_reward/std": 1.8932357668876647, "step": 17250 }, { "clip_ratio/high_max": 0.002296837466322662, "clip_ratio/high_mean": 0.0011618656467665763, "clip_ratio/low_mean": 0.0008674596034931407, "clip_ratio/low_min": 0.00024147887937336538, "clip_ratio/region_mean": 0.0020293252552316212, "completions/clipped_ratio": 0.0, "completions/max_length": 680.5, "completions/max_terminated_length": 680.5, "completions/mean_length": 210.15313110351562, "completions/mean_terminated_length": 210.15313110351562, "completions/min_length": 53.4, "completions/min_terminated_length": 53.4, "entropy": 0.03420282198737065, "epoch": 10.575451484542393, "frac_reward_zero_std": 0.13750000149011612, "grad_norm": 0.0283203125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1768986319.0, "reward": 7.690030813217163, "reward_std": 0.32197879552841185, "rewards/combined_reward/mean": 7.690030479431153, "rewards/combined_reward/std": 2.5868982076644897, "step": 17280 }, { "clip_ratio/high_max": 0.0019526185327170726, "clip_ratio/high_mean": 0.001153846818730623, "clip_ratio/low_mean": 0.0007178072732737443, "clip_ratio/low_min": 0.00022832546310382895, "clip_ratio/region_mean": 0.0018716540856985376, "completions/clipped_ratio": 0.0, "completions/max_length": 557.8, "completions/max_terminated_length": 557.8, "completions/mean_length": 217.83828582763672, "completions/mean_terminated_length": 217.83828582763672, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.027671538355449835, "epoch": 10.59381695745332, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1772122978.0, "reward": 8.065613889694214, "reward_std": 0.36370477378368377, "rewards/combined_reward/mean": 8.065613746643066, "rewards/combined_reward/std": 2.058456552028656, "step": 17310 }, { "clip_ratio/high_max": 0.0018251826123256857, "clip_ratio/high_mean": 0.0010145314258503883, "clip_ratio/low_mean": 0.0007941046023915987, "clip_ratio/low_min": 0.00019053855721722356, "clip_ratio/region_mean": 0.0018086360248465402, "completions/clipped_ratio": 0.0, "completions/max_length": 540.9, "completions/max_terminated_length": 540.9, "completions/mean_length": 202.23906860351562, "completions/mean_terminated_length": 202.23906860351562, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.028536677702019612, "epoch": 10.612182430364248, "frac_reward_zero_std": 0.14583333507180213, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1775199736.0, "reward": 8.02160086631775, "reward_std": 0.36863021552562714, "rewards/combined_reward/mean": 8.021600532531739, "rewards/combined_reward/std": 2.1978826999664305, "step": 17340 }, { "clip_ratio/high_max": 0.0018141220411052928, "clip_ratio/high_mean": 0.0010219699587954286, "clip_ratio/low_mean": 0.0007786502482000894, "clip_ratio/low_min": 0.00027080149617783416, "clip_ratio/region_mean": 0.0018006202042064009, "completions/clipped_ratio": 0.0, "completions/max_length": 686.4, "completions/max_terminated_length": 686.4, "completions/mean_length": 205.59114990234374, "completions/mean_terminated_length": 205.59114990234374, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.02913272922548155, "epoch": 10.630547903275176, "frac_reward_zero_std": 0.14583333544433116, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1778289366.0, "reward": 8.278211832046509, "reward_std": 0.2930243879556656, "rewards/combined_reward/mean": 8.278211688995361, "rewards/combined_reward/std": 2.1317807912826536, "step": 17370 }, { "clip_ratio/high_max": 0.0022768398596478318, "clip_ratio/high_mean": 0.0012756745013272545, "clip_ratio/low_mean": 0.000855701931504882, "clip_ratio/low_min": 0.00027320828400358247, "clip_ratio/region_mean": 0.0021313764542962113, "completions/clipped_ratio": 0.0, "completions/max_length": 549.5, "completions/max_terminated_length": 549.5, "completions/mean_length": 208.13099517822266, "completions/mean_terminated_length": 208.13099517822266, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03097676799322168, "epoch": 10.648913376186103, "frac_reward_zero_std": 0.18750000298023223, "grad_norm": 0.007232666015625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1781388749.0, "reward": 7.993301200866699, "reward_std": 0.3392434626817703, "rewards/combined_reward/mean": 7.993300819396973, "rewards/combined_reward/std": 2.249750316143036, "step": 17400 }, { "clip_ratio/high_max": 0.002099235933565069, "clip_ratio/high_mean": 0.0011854473689406101, "clip_ratio/low_mean": 0.0006682646554812284, "clip_ratio/low_min": 0.0001969454339511382, "clip_ratio/region_mean": 0.0018537120316371631, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 204.66615142822266, "completions/mean_terminated_length": 204.66615142822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.030627268281144402, "epoch": 10.66727884909703, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1784474827.0, "reward": 7.836921262741089, "reward_std": 0.3803052231669426, "rewards/combined_reward/mean": 7.836921167373657, "rewards/combined_reward/std": 2.2885485529899596, "step": 17430 }, { "clip_ratio/high_max": 0.0015931190787038456, "clip_ratio/high_mean": 0.0009259592505259206, "clip_ratio/low_mean": 0.0007701011497071401, "clip_ratio/low_min": 0.00025970745385469246, "clip_ratio/region_mean": 0.0016960603956249543, "completions/clipped_ratio": 0.0, "completions/max_length": 517.5, "completions/max_terminated_length": 517.5, "completions/mean_length": 195.21458892822267, "completions/mean_terminated_length": 195.21458892822267, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.02725476415362209, "epoch": 10.685644322007958, "frac_reward_zero_std": 0.2041666716337204, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1787524611.0, "reward": 8.12702112197876, "reward_std": 0.29104786962270734, "rewards/combined_reward/mean": 8.127020692825317, "rewards/combined_reward/std": 2.165747117996216, "step": 17460 }, { "clip_ratio/high_max": 0.002357443609798793, "clip_ratio/high_mean": 0.0012418250743697476, "clip_ratio/low_mean": 0.0007130672845960361, "clip_ratio/low_min": 0.00018981162284035236, "clip_ratio/region_mean": 0.0019548923373804428, "completions/clipped_ratio": 0.0, "completions/max_length": 533.2, "completions/max_terminated_length": 533.2, "completions/mean_length": 199.59844360351562, "completions/mean_terminated_length": 199.59844360351562, "completions/min_length": 47.6, "completions/min_terminated_length": 47.6, "entropy": 0.029228191605458656, "epoch": 10.704009794918885, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.004547119140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1790591229.0, "reward": 8.191547441482545, "reward_std": 0.3117309495806694, "rewards/combined_reward/mean": 8.191547012329101, "rewards/combined_reward/std": 2.0829647183418274, "step": 17490 }, { "clip_ratio/high_max": 0.0018107397510902957, "clip_ratio/high_mean": 0.0009930096137395594, "clip_ratio/low_mean": 0.0008451458411097216, "clip_ratio/low_min": 0.00022759520361432807, "clip_ratio/region_mean": 0.0018381554451480041, "completions/clipped_ratio": 0.0, "completions/max_length": 574.3, "completions/max_terminated_length": 574.3, "completions/mean_length": 221.0132858276367, "completions/mean_terminated_length": 221.0132858276367, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.028787831139440338, "epoch": 10.722375267829813, "frac_reward_zero_std": 0.10416666902601719, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1793740080.0, "reward": 8.171651458740234, "reward_std": 0.33292798399925233, "rewards/combined_reward/mean": 8.17165150642395, "rewards/combined_reward/std": 2.1267971992492676, "step": 17520 }, { "clip_ratio/high_max": 0.0020021805491220825, "clip_ratio/high_mean": 0.0011405573823140003, "clip_ratio/low_mean": 0.0008323623546554397, "clip_ratio/low_min": 0.00027188568686445553, "clip_ratio/region_mean": 0.0019729197418200785, "completions/clipped_ratio": 0.0, "completions/max_length": 580.4, "completions/max_terminated_length": 580.4, "completions/mean_length": 205.99662017822266, "completions/mean_terminated_length": 205.99662017822266, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.028021884026626746, "epoch": 10.74074074074074, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1796831267.0, "reward": 8.048483037948609, "reward_std": 0.3616132318973541, "rewards/combined_reward/mean": 8.048483085632324, "rewards/combined_reward/std": 2.4137576103210447, "step": 17550 }, { "clip_ratio/high_max": 0.0018017591695146014, "clip_ratio/high_mean": 0.001044235428101577, "clip_ratio/low_mean": 0.0007297882536173953, "clip_ratio/low_min": 0.00021025629296976452, "clip_ratio/region_mean": 0.0017740236944518983, "completions/clipped_ratio": 0.0, "completions/max_length": 595.9, "completions/max_terminated_length": 595.9, "completions/mean_length": 203.6791732788086, "completions/mean_terminated_length": 203.6791732788086, "completions/min_length": 50.4, "completions/min_terminated_length": 50.4, "entropy": 0.02872396114592751, "epoch": 10.759106213651668, "frac_reward_zero_std": 0.1375000037252903, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1799913555.0, "reward": 8.11188325881958, "reward_std": 0.31303083896636963, "rewards/combined_reward/mean": 8.111883163452148, "rewards/combined_reward/std": 2.1496004939079283, "step": 17580 }, { "clip_ratio/high_max": 0.002041918602481019, "clip_ratio/high_mean": 0.0011776423829966613, "clip_ratio/low_mean": 0.0007001179987128125, "clip_ratio/low_min": 0.00021510016449610703, "clip_ratio/region_mean": 0.00187776038122441, "completions/clipped_ratio": 0.0, "completions/max_length": 577.3, "completions/max_terminated_length": 577.3, "completions/mean_length": 214.05547637939452, "completions/mean_terminated_length": 214.05547637939452, "completions/min_length": 49.4, "completions/min_terminated_length": 49.4, "entropy": 0.028862663156663377, "epoch": 10.777471686562595, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1803035688.0, "reward": 8.202723169326783, "reward_std": 0.3020033657550812, "rewards/combined_reward/mean": 8.202722978591918, "rewards/combined_reward/std": 2.1649515271186828, "step": 17610 }, { "clip_ratio/high_max": 0.0019003531041865548, "clip_ratio/high_mean": 0.0010851434043919047, "clip_ratio/low_mean": 0.0008764148666765929, "clip_ratio/low_min": 0.00024265414710195426, "clip_ratio/region_mean": 0.001961558276767998, "completions/clipped_ratio": 0.0, "completions/max_length": 576.5, "completions/max_terminated_length": 576.5, "completions/mean_length": 202.6627685546875, "completions/mean_terminated_length": 202.6627685546875, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.03041901659841339, "epoch": 10.795837159473523, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1806114073.0, "reward": 8.106135129928589, "reward_std": 0.2737629435956478, "rewards/combined_reward/mean": 8.10613489151001, "rewards/combined_reward/std": 2.204833483695984, "step": 17640 }, { "clip_ratio/high_max": 0.0021953103714622557, "clip_ratio/high_mean": 0.0012495196194019325, "clip_ratio/low_mean": 0.0009100286394338278, "clip_ratio/low_min": 0.00027597740045166576, "clip_ratio/region_mean": 0.002159548264656526, "completions/clipped_ratio": 0.0, "completions/max_length": 545.9, "completions/max_terminated_length": 545.9, "completions/mean_length": 195.8768280029297, "completions/mean_terminated_length": 195.8768280029297, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03321255052772661, "epoch": 10.81420263238445, "frac_reward_zero_std": 0.21250000707805156, "grad_norm": 0.00738525390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1809166400.0, "reward": 7.8820885181427, "reward_std": 0.28459183126688004, "rewards/combined_reward/mean": 7.882088470458984, "rewards/combined_reward/std": 2.3806835174560548, "step": 17670 }, { "clip_ratio/high_max": 0.001992406033241423, "clip_ratio/high_mean": 0.0011617084925092057, "clip_ratio/low_mean": 0.0007461096101906151, "clip_ratio/low_min": 0.00018265523394802585, "clip_ratio/region_mean": 0.0019078180999107038, "completions/clipped_ratio": 0.0, "completions/max_length": 655.3, "completions/max_terminated_length": 655.3, "completions/mean_length": 208.63880767822266, "completions/mean_terminated_length": 208.63880767822266, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.029292868512372176, "epoch": 10.832568105295378, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.023681640625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1812267733.0, "reward": 8.236725997924804, "reward_std": 0.35511062741279603, "rewards/combined_reward/mean": 8.236725807189941, "rewards/combined_reward/std": 2.137716794013977, "step": 17700 }, { "clip_ratio/high_max": 0.002284561700556272, "clip_ratio/high_mean": 0.001216808136086911, "clip_ratio/low_mean": 0.0007346427705973231, "clip_ratio/low_min": 0.00017583935696166008, "clip_ratio/region_mean": 0.001951450922448809, "completions/clipped_ratio": 0.0, "completions/max_length": 581.3, "completions/max_terminated_length": 581.3, "completions/mean_length": 201.4953186035156, "completions/mean_terminated_length": 201.4953186035156, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.0311274618220826, "epoch": 10.850933578206305, "frac_reward_zero_std": 0.20000000447034835, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 1815341635.0, "reward": 8.187397956848145, "reward_std": 0.32230047434568404, "rewards/combined_reward/mean": 8.187397575378418, "rewards/combined_reward/std": 2.201090896129608, "step": 17730 }, { "clip_ratio/high_max": 0.0021863082268585763, "clip_ratio/high_mean": 0.0012702252742504546, "clip_ratio/low_mean": 0.0007292853802330986, "clip_ratio/low_min": 0.00017355420786770991, "clip_ratio/region_mean": 0.001999510651512537, "completions/clipped_ratio": 0.0, "completions/max_length": 626.2, "completions/max_terminated_length": 626.2, "completions/mean_length": 208.58568115234374, "completions/mean_terminated_length": 208.58568115234374, "completions/min_length": 53.7, "completions/min_terminated_length": 53.7, "entropy": 0.02964238974576195, "epoch": 10.869299051117233, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1818442764.0, "reward": 8.372621774673462, "reward_std": 0.31571104303002356, "rewards/combined_reward/mean": 8.3726215839386, "rewards/combined_reward/std": 1.8344500958919525, "step": 17760 }, { "clip_ratio/high_max": 0.0018283013089482362, "clip_ratio/high_mean": 0.0010983430067426525, "clip_ratio/low_mean": 0.0007067843875120161, "clip_ratio/low_min": 0.00024311533391786118, "clip_ratio/region_mean": 0.0018051274246924246, "completions/clipped_ratio": 0.0, "completions/max_length": 543.3, "completions/max_terminated_length": 543.3, "completions/mean_length": 210.0367218017578, "completions/mean_terminated_length": 210.0367218017578, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.028621966081360976, "epoch": 10.88766452402816, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.0098876953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1821549465.0, "reward": 8.172498989105225, "reward_std": 0.3426977932453156, "rewards/combined_reward/mean": 8.172498750686646, "rewards/combined_reward/std": 2.162398874759674, "step": 17790 }, { "clip_ratio/high_max": 0.0021148509801908705, "clip_ratio/high_mean": 0.0011868311229288034, "clip_ratio/low_mean": 0.000773829849640606, "clip_ratio/low_min": 0.00021291964076226578, "clip_ratio/region_mean": 0.0019606609566835687, "completions/clipped_ratio": 0.0, "completions/max_length": 608.8, "completions/max_terminated_length": 608.8, "completions/mean_length": 208.0148498535156, "completions/mean_terminated_length": 208.0148498535156, "completions/min_length": 45.7, "completions/min_terminated_length": 45.7, "entropy": 0.028891098328555624, "epoch": 10.906029996939088, "frac_reward_zero_std": 0.16666667275130748, "grad_norm": 0.007415771484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1824648402.0, "reward": 8.095994091033935, "reward_std": 0.3016999587416649, "rewards/combined_reward/mean": 8.095993518829346, "rewards/combined_reward/std": 1.9228736877441406, "step": 17820 }, { "clip_ratio/high_max": 0.0016941533249337226, "clip_ratio/high_mean": 0.0009609856587606676, "clip_ratio/low_mean": 0.0008037277600427236, "clip_ratio/low_min": 0.00020835682710943122, "clip_ratio/region_mean": 0.00176471341959162, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 193.45859985351564, "completions/mean_terminated_length": 193.45859985351564, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.03080838144135972, "epoch": 10.924395469850015, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.00714111328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1827691443.0, "reward": 8.241412448883057, "reward_std": 0.2575531557202339, "rewards/combined_reward/mean": 8.241412115097045, "rewards/combined_reward/std": 2.0479980051517486, "step": 17850 }, { "clip_ratio/high_max": 0.0017071129860899723, "clip_ratio/high_mean": 0.0010787959468871121, "clip_ratio/low_mean": 0.0008051686256294488, "clip_ratio/low_min": 0.0002446351725666318, "clip_ratio/region_mean": 0.0018839645587528746, "completions/clipped_ratio": 0.0, "completions/max_length": 611.9, "completions/max_terminated_length": 611.9, "completions/mean_length": 202.73437957763673, "completions/mean_terminated_length": 202.73437957763673, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.029117555900787313, "epoch": 10.942760942760943, "frac_reward_zero_std": 0.23750000596046447, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1830770103.0, "reward": 7.853962707519531, "reward_std": 0.2606726631522179, "rewards/combined_reward/mean": 7.853962182998657, "rewards/combined_reward/std": 2.57448867559433, "step": 17880 }, { "clip_ratio/high_max": 0.001966498820305181, "clip_ratio/high_mean": 0.0011634282207523938, "clip_ratio/low_mean": 0.0007620542091899551, "clip_ratio/low_min": 0.0002185996022793309, "clip_ratio/region_mean": 0.0019254824408562853, "completions/clipped_ratio": 0.0, "completions/max_length": 556.2, "completions/max_terminated_length": 556.2, "completions/mean_length": 201.21693115234376, "completions/mean_terminated_length": 201.21693115234376, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.028221359398836892, "epoch": 10.96112641567187, "frac_reward_zero_std": 0.2041666753590107, "grad_norm": 0.006103515625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1833842936.0, "reward": 7.9558210372924805, "reward_std": 0.31375248432159425, "rewards/combined_reward/mean": 7.955820894241333, "rewards/combined_reward/std": 2.389954161643982, "step": 17910 }, { "clip_ratio/high_max": 0.0015855436176934745, "clip_ratio/high_mean": 0.0008794090164883528, "clip_ratio/low_mean": 0.0007150313476207278, "clip_ratio/low_min": 0.00020383438653273818, "clip_ratio/region_mean": 0.0015944403625326232, "completions/clipped_ratio": 0.0, "completions/max_length": 650.8, "completions/max_terminated_length": 650.8, "completions/mean_length": 199.10443267822265, "completions/mean_terminated_length": 199.10443267822265, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.025060154769259196, "epoch": 10.979491888582798, "frac_reward_zero_std": 0.1625000026077032, "grad_norm": 0.0107421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 1836907657.0, "reward": 8.079147243499756, "reward_std": 0.2917182069271803, "rewards/combined_reward/mean": 8.079146766662598, "rewards/combined_reward/std": 2.122173011302948, "step": 17940 }, { "clip_ratio/high_max": 0.0018680448481366815, "clip_ratio/high_mean": 0.0010608991848130244, "clip_ratio/low_mean": 0.0008369658433366566, "clip_ratio/low_min": 0.0002994571293432576, "clip_ratio/region_mean": 0.0018978650371233622, "completions/clipped_ratio": 0.0, "completions/max_length": 578.8, "completions/max_terminated_length": 578.8, "completions/mean_length": 203.11015930175782, "completions/mean_terminated_length": 203.11015930175782, "completions/min_length": 49.4, "completions/min_terminated_length": 49.4, "entropy": 0.025259835133329033, "epoch": 10.997857361493725, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1839987760.0, "reward": 8.15322012901306, "reward_std": 0.30352861136198045, "rewards/combined_reward/mean": 8.153219985961915, "rewards/combined_reward/std": 2.2719157338142395, "step": 17970 }, { "clip_ratio/high_max": 0.002281382041086336, "clip_ratio/high_mean": 0.0013118775885410943, "clip_ratio/low_mean": 0.0008348321858707617, "clip_ratio/low_min": 0.00019624198049299235, "clip_ratio/region_mean": 0.0021467097843119293, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 201.49036865234376, "completions/mean_terminated_length": 201.49036865234376, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02633849537519158, "epoch": 11.016528925619834, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.01025390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1843061643.0, "reward": 8.126143836975098, "reward_std": 0.29920888096094134, "rewards/combined_reward/mean": 8.126143598556519, "rewards/combined_reward/std": 2.230824810266495, "step": 18000 }, { "clip_ratio/high_max": 0.002590834513345423, "clip_ratio/high_mean": 0.0015275886515155434, "clip_ratio/low_mean": 0.000993561819450406, "clip_ratio/low_min": 0.00032658754450191435, "clip_ratio/region_mean": 0.002521150434525528, "completions/clipped_ratio": 0.0, "completions/max_length": 605.4, "completions/max_terminated_length": 605.4, "completions/mean_length": 206.3072982788086, "completions/mean_terminated_length": 206.3072982788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.031160873423020045, "epoch": 11.034894398530762, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1846154023.0, "reward": 8.009607791900635, "reward_std": 0.3271432861685753, "rewards/combined_reward/mean": 8.009607601165772, "rewards/combined_reward/std": 2.285069453716278, "step": 18030 }, { "clip_ratio/high_max": 0.0024206875100693044, "clip_ratio/high_mean": 0.001538295132437876, "clip_ratio/low_mean": 0.0009370715387679714, "clip_ratio/low_min": 0.00028119403553622155, "clip_ratio/region_mean": 0.002475366673267369, "completions/clipped_ratio": 0.0, "completions/max_length": 558.2, "completions/max_terminated_length": 558.2, "completions/mean_length": 205.3937561035156, "completions/mean_terminated_length": 205.3937561035156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.0281584898630778, "epoch": 11.05325987144169, "frac_reward_zero_std": 0.13333333656191826, "grad_norm": 0.0050048828125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1849242895.0, "reward": 8.162663316726684, "reward_std": 0.2801635578274727, "rewards/combined_reward/mean": 8.162663078308105, "rewards/combined_reward/std": 2.0512362718582153, "step": 18060 }, { "clip_ratio/high_max": 0.0022623488621320575, "clip_ratio/high_mean": 0.0013101679965378328, "clip_ratio/low_mean": 0.0010033244961960008, "clip_ratio/low_min": 0.00024347441115727027, "clip_ratio/region_mean": 0.002313492500494855, "completions/clipped_ratio": 0.0, "completions/max_length": 502.9, "completions/max_terminated_length": 502.9, "completions/mean_length": 202.50130767822264, "completions/mean_terminated_length": 202.50130767822264, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.027775541273877026, "epoch": 11.071625344352617, "frac_reward_zero_std": 0.11666666828095913, "grad_norm": 0.015380859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 1852320660.0, "reward": 8.291706466674805, "reward_std": 0.3694057434797287, "rewards/combined_reward/mean": 8.291706132888795, "rewards/combined_reward/std": 2.1375715494155885, "step": 18090 }, { "clip_ratio/high_max": 0.0023171142934491702, "clip_ratio/high_mean": 0.0013848419656521096, "clip_ratio/low_mean": 0.0008789655402021405, "clip_ratio/low_min": 0.00024773788779081465, "clip_ratio/region_mean": 0.0022638075233165486, "completions/clipped_ratio": 0.0, "completions/max_length": 555.3, "completions/max_terminated_length": 555.3, "completions/mean_length": 203.18021087646486, "completions/mean_terminated_length": 203.18021087646486, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02949258186854422, "epoch": 11.089990817263544, "frac_reward_zero_std": 0.14583333507180213, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1855401032.0, "reward": 8.075627708435059, "reward_std": 0.30661137104034425, "rewards/combined_reward/mean": 8.075627517700195, "rewards/combined_reward/std": 2.054232180118561, "step": 18120 }, { "clip_ratio/high_max": 0.0023165820270757346, "clip_ratio/high_mean": 0.0014804887521070972, "clip_ratio/low_mean": 0.0008597936928102475, "clip_ratio/low_min": 0.0002648840347925822, "clip_ratio/region_mean": 0.0023402824304260624, "completions/clipped_ratio": 0.0, "completions/max_length": 807.3, "completions/max_terminated_length": 807.3, "completions/mean_length": 207.1343795776367, "completions/mean_terminated_length": 207.1343795776367, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.030517315538600088, "epoch": 11.108356290174472, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1858496588.0, "reward": 7.87665319442749, "reward_std": 0.321976038813591, "rewards/combined_reward/mean": 7.876653146743775, "rewards/combined_reward/std": 2.3524207353591917, "step": 18150 }, { "clip_ratio/high_max": 0.002680679912737105, "clip_ratio/high_mean": 0.001509564064569228, "clip_ratio/low_mean": 0.0010793134876924645, "clip_ratio/low_min": 0.0002655804928508587, "clip_ratio/region_mean": 0.0025888775926432572, "completions/clipped_ratio": 0.0, "completions/max_length": 683.1, "completions/max_terminated_length": 683.1, "completions/mean_length": 202.3065185546875, "completions/mean_terminated_length": 202.3065185546875, "completions/min_length": 41.1, "completions/min_terminated_length": 41.1, "entropy": 0.032706758053973314, "epoch": 11.1267217630854, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1861573605.0, "reward": 8.000605869293214, "reward_std": 0.2998155653476715, "rewards/combined_reward/mean": 8.000605726242066, "rewards/combined_reward/std": 2.455345630645752, "step": 18180 }, { "clip_ratio/high_max": 0.0026120224501937628, "clip_ratio/high_mean": 0.0013962498220886725, "clip_ratio/low_mean": 0.0009792830705312857, "clip_ratio/low_min": 0.0002618010492975979, "clip_ratio/region_mean": 0.0023755329030488307, "completions/clipped_ratio": 0.0, "completions/max_length": 649.6, "completions/max_terminated_length": 649.6, "completions/mean_length": 196.13099365234376, "completions/mean_terminated_length": 196.13099365234376, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029724345185483496, "epoch": 11.145087235996327, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.0037689208984375, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 1864626908.0, "reward": 7.96673698425293, "reward_std": 0.31017170250415804, "rewards/combined_reward/mean": 7.9667370319366455, "rewards/combined_reward/std": 2.333874762058258, "step": 18210 }, { "clip_ratio/high_max": 0.0024275368346328226, "clip_ratio/high_mean": 0.0013034386731305859, "clip_ratio/low_mean": 0.001030813176233399, "clip_ratio/low_min": 0.00035709483684816706, "clip_ratio/region_mean": 0.002334251827172314, "completions/clipped_ratio": 0.0, "completions/max_length": 528.3, "completions/max_terminated_length": 528.3, "completions/mean_length": 206.0875030517578, "completions/mean_terminated_length": 206.0875030517578, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.029024180738876262, "epoch": 11.163452708907254, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1867718444.0, "reward": 8.169423151016236, "reward_std": 0.31478137969970704, "rewards/combined_reward/mean": 8.169422912597657, "rewards/combined_reward/std": 2.05205694437027, "step": 18240 }, { "clip_ratio/high_max": 0.0023648039447531724, "clip_ratio/high_mean": 0.0013245427452299432, "clip_ratio/low_mean": 0.0008858907759228411, "clip_ratio/low_min": 0.0002379243572552999, "clip_ratio/region_mean": 0.002210433503690486, "completions/clipped_ratio": 0.0, "completions/max_length": 532.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 201.5721405029297, "completions/mean_terminated_length": 201.5721405029297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.02536132667058458, "epoch": 11.181818181818182, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1870792641.0, "reward": 8.089344549179078, "reward_std": 0.2353600338101387, "rewards/combined_reward/mean": 8.089344358444214, "rewards/combined_reward/std": 2.3315412223339083, "step": 18270 }, { "clip_ratio/high_max": 0.002715250660548918, "clip_ratio/high_mean": 0.0015619385955991068, "clip_ratio/low_mean": 0.0009392848526355616, "clip_ratio/low_min": 0.00022731853290072952, "clip_ratio/region_mean": 0.0025012234308330033, "completions/clipped_ratio": 0.0, "completions/max_length": 547.8, "completions/max_terminated_length": 547.8, "completions/mean_length": 196.62708740234376, "completions/mean_terminated_length": 196.62708740234376, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.03004926724048952, "epoch": 11.20018365472911, "frac_reward_zero_std": 0.2333333395421505, "grad_norm": 0.026123046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1873847849.0, "reward": 8.057130575180054, "reward_std": 0.28794110119342803, "rewards/combined_reward/mean": 8.057130241394043, "rewards/combined_reward/std": 2.228439688682556, "step": 18300 }, { "clip_ratio/high_max": 0.0020554328133584932, "clip_ratio/high_mean": 0.0011651525556468792, "clip_ratio/low_mean": 0.0008391059421531585, "clip_ratio/low_min": 0.00013429225582513028, "clip_ratio/region_mean": 0.0020042584964054793, "completions/clipped_ratio": 0.0, "completions/max_length": 673.5, "completions/max_terminated_length": 673.5, "completions/mean_length": 200.04141235351562, "completions/mean_terminated_length": 200.04141235351562, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.026018949310916164, "epoch": 11.218549127640037, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1876916168.0, "reward": 7.975059604644775, "reward_std": 0.28449649810791017, "rewards/combined_reward/mean": 7.975059270858765, "rewards/combined_reward/std": 2.4898298740386964, "step": 18330 }, { "clip_ratio/high_max": 0.002651784634508658, "clip_ratio/high_mean": 0.0014523027297400405, "clip_ratio/low_mean": 0.0010830318635271397, "clip_ratio/low_min": 0.0003005612901082107, "clip_ratio/region_mean": 0.0025353345825957756, "completions/clipped_ratio": 0.0, "completions/max_length": 584.9, "completions/max_terminated_length": 584.9, "completions/mean_length": 194.29297485351563, "completions/mean_terminated_length": 194.29297485351563, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.03214526839243869, "epoch": 11.236914600550964, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.004150390625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1879962413.0, "reward": 8.08599820137024, "reward_std": 0.28808602392673494, "rewards/combined_reward/mean": 8.085997819900513, "rewards/combined_reward/std": 2.116376888751984, "step": 18360 }, { "clip_ratio/high_max": 0.0018974004259992702, "clip_ratio/high_mean": 0.0011832228890853002, "clip_ratio/low_mean": 0.0008923486694887591, "clip_ratio/low_min": 0.00025095050417197246, "clip_ratio/region_mean": 0.002075571571670783, "completions/clipped_ratio": 0.0, "completions/max_length": 698.6, "completions/max_terminated_length": 698.6, "completions/mean_length": 204.84740142822267, "completions/mean_terminated_length": 204.84740142822267, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.02603357257321477, "epoch": 11.255280073461892, "frac_reward_zero_std": 0.1291666690260172, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1883049187.0, "reward": 8.178902959823608, "reward_std": 0.3065895766019821, "rewards/combined_reward/mean": 8.178902626037598, "rewards/combined_reward/std": 2.1261917114257813, "step": 18390 }, { "clip_ratio/high_max": 0.0018488860565412323, "clip_ratio/high_mean": 0.0011010484196958715, "clip_ratio/low_mean": 0.0007540732587585808, "clip_ratio/low_min": 0.0002445238878863165, "clip_ratio/region_mean": 0.001855121697008144, "completions/clipped_ratio": 0.0, "completions/max_length": 578.4285714285714, "completions/max_terminated_length": 578.4285714285714, "completions/mean_length": 206.8024640764509, "completions/mean_terminated_length": 206.8024640764509, "completions/min_length": 37.714285714285715, "completions/min_terminated_length": 37.714285714285715, "entropy": 0.02944043402094394, "epoch": 11.273033363942455, "frac_reward_zero_std": 0.13690476438828877, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1886128618.0, "reward": 8.124837534768242, "reward_std": 0.364823294537408, "rewards/combined_reward/mean": 8.124837398529053, "rewards/combined_reward/std": 2.173139844621931, "step": 18420 }, { "clip_ratio/high_max": 0.0017911215181811713, "clip_ratio/high_mean": 0.001048277671240309, "clip_ratio/low_mean": 0.0007342221503980303, "clip_ratio/low_min": 0.0002380077651954101, "clip_ratio/region_mean": 0.0017824998411621587, "completions/clipped_ratio": 0.0, "completions/max_length": 640.2, "completions/max_terminated_length": 640.2, "completions/mean_length": 206.58307800292968, "completions/mean_terminated_length": 206.58307800292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02951746042817831, "epoch": 11.291398836853382, "frac_reward_zero_std": 0.18750000223517418, "grad_norm": 0.024169921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1889222057.0, "reward": 8.048280906677245, "reward_std": 0.28340347707271574, "rewards/combined_reward/mean": 8.048280572891235, "rewards/combined_reward/std": 2.1546345591545104, "step": 18450 }, { "clip_ratio/high_max": 0.0018356537764096478, "clip_ratio/high_mean": 0.0010517293008888372, "clip_ratio/low_mean": 0.0006840426845883485, "clip_ratio/low_min": 0.00022396229154158693, "clip_ratio/region_mean": 0.0017357719887513667, "completions/clipped_ratio": 0.0, "completions/max_length": 574.1, "completions/max_terminated_length": 574.1, "completions/mean_length": 205.48151702880858, "completions/mean_terminated_length": 205.48151702880858, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028403436935817202, "epoch": 11.30976430976431, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 1892311266.0, "reward": 7.902102041244507, "reward_std": 0.3218313202261925, "rewards/combined_reward/mean": 7.902101659774781, "rewards/combined_reward/std": 2.2809115290641784, "step": 18480 }, { "clip_ratio/high_max": 0.00213057005845864, "clip_ratio/high_mean": 0.0012999142145417864, "clip_ratio/low_mean": 0.0007535479007250008, "clip_ratio/low_min": 0.0001951164939479592, "clip_ratio/region_mean": 0.002053462102776393, "completions/clipped_ratio": 0.0, "completions/max_length": 630.5, "completions/max_terminated_length": 630.5, "completions/mean_length": 200.9862030029297, "completions/mean_terminated_length": 200.9862030029297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027995704424877962, "epoch": 11.328129782675237, "frac_reward_zero_std": 0.2083333373069763, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1895383213.0, "reward": 7.945948553085327, "reward_std": 0.2867796115577221, "rewards/combined_reward/mean": 7.945948457717895, "rewards/combined_reward/std": 2.386673891544342, "step": 18510 }, { "clip_ratio/high_max": 0.002309750364899325, "clip_ratio/high_mean": 0.0013079148637189064, "clip_ratio/low_mean": 0.0009234508805699685, "clip_ratio/low_min": 0.00021860368603180783, "clip_ratio/region_mean": 0.002231365744955838, "completions/clipped_ratio": 0.0, "completions/max_length": 562.1, "completions/max_terminated_length": 562.1, "completions/mean_length": 202.7487030029297, "completions/mean_terminated_length": 202.7487030029297, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.03390745134092867, "epoch": 11.346495255586165, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1898461928.0, "reward": 8.171932649612426, "reward_std": 0.2755375638604164, "rewards/combined_reward/mean": 8.171932601928711, "rewards/combined_reward/std": 2.030192291736603, "step": 18540 }, { "clip_ratio/high_max": 0.0020132620610335532, "clip_ratio/high_mean": 0.0011271430853715477, "clip_ratio/low_mean": 0.0008389794955292018, "clip_ratio/low_min": 0.0002580276426063695, "clip_ratio/region_mean": 0.0019661225659850365, "completions/clipped_ratio": 0.0, "completions/max_length": 607.2, "completions/max_terminated_length": 607.2, "completions/mean_length": 209.8713623046875, "completions/mean_terminated_length": 209.8713623046875, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.030413024492251375, "epoch": 11.364860728497092, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1901567994.0, "reward": 8.184772396087647, "reward_std": 0.33019678890705106, "rewards/combined_reward/mean": 8.18477201461792, "rewards/combined_reward/std": 2.3513732671737673, "step": 18570 }, { "clip_ratio/high_max": 0.0018095491349716516, "clip_ratio/high_mean": 0.001033293483487796, "clip_ratio/low_mean": 0.0006970926081218446, "clip_ratio/low_min": 0.00015459844459352705, "clip_ratio/region_mean": 0.0017303860882141938, "completions/clipped_ratio": 0.0, "completions/max_length": 619.7, "completions/max_terminated_length": 619.7, "completions/mean_length": 206.1242233276367, "completions/mean_terminated_length": 206.1242233276367, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02854483889726301, "epoch": 11.38322620140802, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1904659671.0, "reward": 7.925138664245606, "reward_std": 0.404203000664711, "rewards/combined_reward/mean": 7.9251384258270265, "rewards/combined_reward/std": 2.3482935786247254, "step": 18600 }, { "clip_ratio/high_max": 0.0021400747287164752, "clip_ratio/high_mean": 0.0012199012949470974, "clip_ratio/low_mean": 0.000724055173365438, "clip_ratio/low_min": 0.00021246623024732494, "clip_ratio/region_mean": 0.0019439564765586206, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 195.4406311035156, "completions/mean_terminated_length": 195.4406311035156, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029021091138323147, "epoch": 11.401591674318947, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1907710323.0, "reward": 8.122975540161132, "reward_std": 0.29924692064523695, "rewards/combined_reward/mean": 8.122975301742553, "rewards/combined_reward/std": 2.1420301675796507, "step": 18630 }, { "clip_ratio/high_max": 0.002173745701050696, "clip_ratio/high_mean": 0.0012482578244089383, "clip_ratio/low_mean": 0.0008396510066328726, "clip_ratio/low_min": 0.0002110856854415033, "clip_ratio/region_mean": 0.0020879088284952254, "completions/clipped_ratio": 0.0, "completions/max_length": 588.7, "completions/max_terminated_length": 588.7, "completions/mean_length": 201.38906555175782, "completions/mean_terminated_length": 201.38906555175782, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.030278626487900813, "epoch": 11.419957147229875, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1910783817.0, "reward": 8.089427995681763, "reward_std": 0.33442931324243547, "rewards/combined_reward/mean": 8.089427757263184, "rewards/combined_reward/std": 2.2930360198020936, "step": 18660 }, { "clip_ratio/high_max": 0.0020146078313700855, "clip_ratio/high_mean": 0.0011404197105245354, "clip_ratio/low_mean": 0.000691295979777351, "clip_ratio/low_min": 0.00018785839856718666, "clip_ratio/region_mean": 0.0018317156951525248, "completions/clipped_ratio": 0.0, "completions/max_length": 509.6, "completions/max_terminated_length": 509.6, "completions/mean_length": 188.74557495117188, "completions/mean_terminated_length": 188.74557495117188, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0321498669218272, "epoch": 11.438322620140802, "frac_reward_zero_std": 0.20833334252238273, "grad_norm": 0.009765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1913808760.0, "reward": 7.9946760654449465, "reward_std": 0.3118949502706528, "rewards/combined_reward/mean": 7.994675827026367, "rewards/combined_reward/std": 2.3388846039772035, "step": 18690 }, { "clip_ratio/high_max": 0.0018691466233576649, "clip_ratio/high_mean": 0.0010495993148651906, "clip_ratio/low_mean": 0.0008580581214725195, "clip_ratio/low_min": 0.0002585756160745708, "clip_ratio/region_mean": 0.0019076574370653057, "completions/clipped_ratio": 0.0, "completions/max_length": 569.9, "completions/max_terminated_length": 569.9, "completions/mean_length": 200.2726608276367, "completions/mean_terminated_length": 200.2726608276367, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029640713582436244, "epoch": 11.45668809305173, "frac_reward_zero_std": 0.13750000149011612, "grad_norm": 0.025634765625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 1916877967.0, "reward": 7.865443229675293, "reward_std": 0.3695697486400604, "rewards/combined_reward/mean": 7.865443181991577, "rewards/combined_reward/std": 2.235736459493637, "step": 18720 }, { "clip_ratio/high_max": 0.0019249773356326235, "clip_ratio/high_mean": 0.0010698227344012898, "clip_ratio/low_mean": 0.0007072666376795193, "clip_ratio/low_min": 0.00016090268230376144, "clip_ratio/region_mean": 0.0017770893789323357, "completions/clipped_ratio": 0.0, "completions/max_length": 603.1, "completions/max_terminated_length": 603.1, "completions/mean_length": 197.32109832763672, "completions/mean_terminated_length": 197.32109832763672, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.026302796990300217, "epoch": 11.475053565962657, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.0025634765625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1919935840.0, "reward": 8.143269395828247, "reward_std": 0.2838803336024284, "rewards/combined_reward/mean": 8.143269157409668, "rewards/combined_reward/std": 2.2529715299606323, "step": 18750 }, { "clip_ratio/high_max": 0.002086095027819586, "clip_ratio/high_mean": 0.0010912485682638361, "clip_ratio/low_mean": 0.0008022112415346783, "clip_ratio/low_min": 0.00023288381586705024, "clip_ratio/region_mean": 0.0018934598192572594, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 200.50104827880858, "completions/mean_terminated_length": 200.50104827880858, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.03228451758623123, "epoch": 11.493419038873585, "frac_reward_zero_std": 0.15000000596046448, "grad_norm": 0.006072998046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1923005924.0, "reward": 8.132836961746216, "reward_std": 0.3115999221801758, "rewards/combined_reward/mean": 8.132836818695068, "rewards/combined_reward/std": 2.2701339602470396, "step": 18780 }, { "clip_ratio/high_max": 0.0018197408110912268, "clip_ratio/high_mean": 0.0010871038720021411, "clip_ratio/low_mean": 0.0007578563305893719, "clip_ratio/low_min": 0.00013970762520330027, "clip_ratio/region_mean": 0.0018449602018032845, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 197.6190155029297, "completions/mean_terminated_length": 197.6190155029297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027742507960647346, "epoch": 11.511784511784512, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 1926064941.0, "reward": 7.755133771896363, "reward_std": 0.3489619717001915, "rewards/combined_reward/mean": 7.755133628845215, "rewards/combined_reward/std": 2.5716392755508424, "step": 18810 }, { "clip_ratio/high_max": 0.001718714410768977, "clip_ratio/high_mean": 0.0009980944776543765, "clip_ratio/low_mean": 0.0007355650263586237, "clip_ratio/low_min": 0.0001855478847573977, "clip_ratio/region_mean": 0.001733659513411112, "completions/clipped_ratio": 0.0, "completions/max_length": 612.5, "completions/max_terminated_length": 612.5, "completions/mean_length": 193.23958892822264, "completions/mean_terminated_length": 193.23958892822264, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.028328694542869924, "epoch": 11.53014998469544, "frac_reward_zero_std": 0.18750000484287738, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 1929107141.0, "reward": 7.891829967498779, "reward_std": 0.2826699808239937, "rewards/combined_reward/mean": 7.8918296813964846, "rewards/combined_reward/std": 2.458651602268219, "step": 18840 }, { "clip_ratio/high_max": 0.0016626413867925293, "clip_ratio/high_mean": 0.0008808480928564677, "clip_ratio/low_mean": 0.0007396033916544791, "clip_ratio/low_min": 0.00019661995732652335, "clip_ratio/region_mean": 0.001620451492878298, "completions/clipped_ratio": 0.0, "completions/max_length": 584.3, "completions/max_terminated_length": 584.3, "completions/mean_length": 195.14115142822266, "completions/mean_terminated_length": 195.14115142822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.026531887419211368, "epoch": 11.548515457606367, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1932156643.0, "reward": 8.23739538192749, "reward_std": 0.2571980461478233, "rewards/combined_reward/mean": 8.23739504814148, "rewards/combined_reward/std": 2.165886926651001, "step": 18870 }, { "clip_ratio/high_max": 0.0020051141970422273, "clip_ratio/high_mean": 0.0010821360870371184, "clip_ratio/low_mean": 0.0008903223679226357, "clip_ratio/low_min": 0.0002573464866145514, "clip_ratio/region_mean": 0.0019724584417417645, "completions/clipped_ratio": 0.0, "completions/max_length": 584.2, "completions/max_terminated_length": 584.2, "completions/mean_length": 208.29141235351562, "completions/mean_terminated_length": 208.29141235351562, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.031191333616152405, "epoch": 11.566880930517295, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1935256642.0, "reward": 7.933788537979126, "reward_std": 0.2970860406756401, "rewards/combined_reward/mean": 7.933788108825683, "rewards/combined_reward/std": 2.241314685344696, "step": 18900 }, { "clip_ratio/high_max": 0.0017075233336072415, "clip_ratio/high_mean": 0.0009777877095378548, "clip_ratio/low_mean": 0.000751185569788504, "clip_ratio/low_min": 0.00016879235893914787, "clip_ratio/region_mean": 0.001728973293696375, "completions/clipped_ratio": 0.0, "completions/max_length": 563.9, "completions/max_terminated_length": 563.9, "completions/mean_length": 208.50833740234376, "completions/mean_terminated_length": 208.50833740234376, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.02866348666138947, "epoch": 11.585246403428222, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.00555419921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1938357474.0, "reward": 7.791168928146362, "reward_std": 0.3695371955633163, "rewards/combined_reward/mean": 7.7911687850952145, "rewards/combined_reward/std": 2.3724981904029847, "step": 18930 }, { "clip_ratio/high_max": 0.001967946846950023, "clip_ratio/high_mean": 0.0011029982398516344, "clip_ratio/low_mean": 0.0007211509640910662, "clip_ratio/low_min": 0.0002243471370699505, "clip_ratio/region_mean": 0.001824149226983233, "completions/clipped_ratio": 0.0, "completions/max_length": 552.2, "completions/max_terminated_length": 552.2, "completions/mean_length": 214.2096405029297, "completions/mean_terminated_length": 214.2096405029297, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.027729808694372575, "epoch": 11.60361187633915, "frac_reward_zero_std": 0.13333333656191826, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1941480199.0, "reward": 8.173706865310669, "reward_std": 0.36743455529212954, "rewards/combined_reward/mean": 8.173706674575806, "rewards/combined_reward/std": 2.0505524277687073, "step": 18960 }, { "clip_ratio/high_max": 0.0017648161485946427, "clip_ratio/high_mean": 0.0009984524508278507, "clip_ratio/low_mean": 0.0006850051033931474, "clip_ratio/low_min": 0.0001197914049650232, "clip_ratio/region_mean": 0.001683457542336934, "completions/clipped_ratio": 0.0, "completions/max_length": 577.4, "completions/max_terminated_length": 577.4, "completions/mean_length": 197.52292022705078, "completions/mean_terminated_length": 197.52292022705078, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.02817059902784725, "epoch": 11.621977349250077, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.006500244140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 1944538847.0, "reward": 8.19272756576538, "reward_std": 0.3210770830512047, "rewards/combined_reward/mean": 8.19272713661194, "rewards/combined_reward/std": 2.0981581807136536, "step": 18990 }, { "clip_ratio/high_max": 0.0019345769062056207, "clip_ratio/high_mean": 0.0010401681949588238, "clip_ratio/low_mean": 0.000876089216156591, "clip_ratio/low_min": 0.0002512788259385464, "clip_ratio/region_mean": 0.0019162573977761592, "completions/clipped_ratio": 0.0, "completions/max_length": 563.8, "completions/max_terminated_length": 563.8, "completions/mean_length": 210.22552795410155, "completions/mean_terminated_length": 210.22552795410155, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.030892535873378316, "epoch": 11.640342822161005, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.01507568359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1947646273.0, "reward": 8.15208420753479, "reward_std": 0.28747256845235825, "rewards/combined_reward/mean": 8.152083921432496, "rewards/combined_reward/std": 2.044968616962433, "step": 19020 }, { "clip_ratio/high_max": 0.001958388096682029, "clip_ratio/high_mean": 0.0011766982015008884, "clip_ratio/low_mean": 0.0008453749913314823, "clip_ratio/low_min": 0.00026533953326482636, "clip_ratio/region_mean": 0.0020220731928323704, "completions/clipped_ratio": 0.0, "completions/max_length": 574.6, "completions/max_terminated_length": 574.6, "completions/mean_length": 206.7447982788086, "completions/mean_terminated_length": 206.7447982788086, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.029553811174506943, "epoch": 11.658708295071932, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1950740333.0, "reward": 8.019474267959595, "reward_std": 0.35539896190166476, "rewards/combined_reward/mean": 8.019473886489868, "rewards/combined_reward/std": 2.363179695606232, "step": 19050 }, { "clip_ratio/high_max": 0.0018138826402719133, "clip_ratio/high_mean": 0.0011116948918546163, "clip_ratio/low_mean": 0.0007853894556319573, "clip_ratio/low_min": 0.00021925495796798106, "clip_ratio/region_mean": 0.0018970843414232756, "completions/clipped_ratio": 0.0, "completions/max_length": 547.1, "completions/max_terminated_length": 547.1, "completions/mean_length": 198.58021392822266, "completions/mean_terminated_length": 198.58021392822266, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.029511913750320674, "epoch": 11.67707376798286, "frac_reward_zero_std": 0.22500000670552253, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1953803041.0, "reward": 7.8933672428131105, "reward_std": 0.360595266520977, "rewards/combined_reward/mean": 7.893367147445678, "rewards/combined_reward/std": 2.2165902853012085, "step": 19080 }, { "clip_ratio/high_max": 0.001974089057572807, "clip_ratio/high_mean": 0.0010804041790834162, "clip_ratio/low_mean": 0.0007061234940920258, "clip_ratio/low_min": 0.00020758203439375696, "clip_ratio/region_mean": 0.0017865276880911552, "completions/clipped_ratio": 0.0, "completions/max_length": 556.1, "completions/max_terminated_length": 556.1, "completions/mean_length": 197.6822952270508, "completions/mean_terminated_length": 197.6822952270508, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.02812282821784417, "epoch": 11.695439240893787, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 1956862301.0, "reward": 8.132514333724975, "reward_std": 0.24692163467407227, "rewards/combined_reward/mean": 8.132514190673827, "rewards/combined_reward/std": 2.178479588031769, "step": 19110 }, { "clip_ratio/high_max": 0.0019629845720677015, "clip_ratio/high_mean": 0.0011002021954482188, "clip_ratio/low_mean": 0.0008173487463257819, "clip_ratio/low_min": 0.00019369062065379694, "clip_ratio/region_mean": 0.0019175509466246392, "completions/clipped_ratio": 0.0, "completions/max_length": 567.7, "completions/max_terminated_length": 567.7, "completions/mean_length": 217.19505767822267, "completions/mean_terminated_length": 217.19505767822267, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.0290052546809117, "epoch": 11.713804713804715, "frac_reward_zero_std": 0.1041666679084301, "grad_norm": 0.05078125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 1959996490.0, "reward": 8.287183046340942, "reward_std": 0.40309927612543106, "rewards/combined_reward/mean": 8.287182521820068, "rewards/combined_reward/std": 1.9200875997543334, "step": 19140 }, { "clip_ratio/high_max": 0.00163217038110209, "clip_ratio/high_mean": 0.000930804717548502, "clip_ratio/low_mean": 0.0008364505223047066, "clip_ratio/low_min": 0.00023436898021221472, "clip_ratio/region_mean": 0.001767255223482304, "completions/clipped_ratio": 0.0, "completions/max_length": 590.3, "completions/max_terminated_length": 590.3, "completions/mean_length": 209.25130920410157, "completions/mean_terminated_length": 209.25130920410157, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.028208483864242833, "epoch": 11.732170186715642, "frac_reward_zero_std": 0.14166667200624944, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 1963100175.0, "reward": 8.011083936691284, "reward_std": 0.3538018137216568, "rewards/combined_reward/mean": 8.011083555221557, "rewards/combined_reward/std": 2.509057569503784, "step": 19170 }, { "clip_ratio/high_max": 0.0019018798130370365, "clip_ratio/high_mean": 0.0010526534152935105, "clip_ratio/low_mean": 0.0006817060663403632, "clip_ratio/low_min": 0.00019664346788583013, "clip_ratio/region_mean": 0.0017343594915776824, "completions/clipped_ratio": 0.0, "completions/max_length": 562.4, "completions/max_terminated_length": 562.4, "completions/mean_length": 203.9401107788086, "completions/mean_terminated_length": 203.9401107788086, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.02631577878880004, "epoch": 11.75053565962657, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 1966183465.0, "reward": 8.173610305786132, "reward_std": 0.34808813631534574, "rewards/combined_reward/mean": 8.173609924316406, "rewards/combined_reward/std": 2.167420965433121, "step": 19200 }, { "clip_ratio/high_max": 0.0021312639000825584, "clip_ratio/high_mean": 0.001261807376310268, "clip_ratio/low_mean": 0.0007643585619613683, "clip_ratio/low_min": 0.00025891570646005373, "clip_ratio/region_mean": 0.002026165953914945, "completions/clipped_ratio": 0.0, "completions/max_length": 543.3, "completions/max_terminated_length": 543.3, "completions/mean_length": 206.24584045410157, "completions/mean_terminated_length": 206.24584045410157, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.031640285688141985, "epoch": 11.768901132537497, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 1969275609.0, "reward": 8.047360754013061, "reward_std": 0.3653345823287964, "rewards/combined_reward/mean": 8.04736042022705, "rewards/combined_reward/std": 2.2868366837501526, "step": 19230 }, { "clip_ratio/high_max": 0.0017876458975176017, "clip_ratio/high_mean": 0.000998353332882592, "clip_ratio/low_mean": 0.000732597990766711, "clip_ratio/low_min": 0.00019173092199101423, "clip_ratio/region_mean": 0.0017309513195262602, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 206.54088897705077, "completions/mean_terminated_length": 206.54088897705077, "completions/min_length": 47.6, "completions/min_terminated_length": 47.6, "entropy": 0.026288906081269183, "epoch": 11.787266605448423, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1972368886.0, "reward": 8.154892873764037, "reward_std": 0.2747855231165886, "rewards/combined_reward/mean": 8.154892539978027, "rewards/combined_reward/std": 2.1798290848731994, "step": 19260 }, { "clip_ratio/high_max": 0.0022445158324747656, "clip_ratio/high_mean": 0.0012106812616063205, "clip_ratio/low_mean": 0.0007855149397073546, "clip_ratio/low_min": 0.00013561403247877024, "clip_ratio/region_mean": 0.0019961962238691437, "completions/clipped_ratio": 0.0, "completions/max_length": 571.9, "completions/max_terminated_length": 571.9, "completions/mean_length": 205.98516082763672, "completions/mean_terminated_length": 205.98516082763672, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03256705032351116, "epoch": 11.805632078359352, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 1975460029.0, "reward": 8.115085363388062, "reward_std": 0.2825535282492638, "rewards/combined_reward/mean": 8.11508493423462, "rewards/combined_reward/std": 2.112857985496521, "step": 19290 }, { "clip_ratio/high_max": 0.0019567862288871157, "clip_ratio/high_mean": 0.0011316045694305407, "clip_ratio/low_mean": 0.0008829579543089494, "clip_ratio/low_min": 0.00024933683792672433, "clip_ratio/region_mean": 0.0020145625181612558, "completions/clipped_ratio": 0.0, "completions/max_length": 553.8, "completions/max_terminated_length": 553.8, "completions/mean_length": 197.35521087646484, "completions/mean_terminated_length": 197.35521087646484, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029777461662888527, "epoch": 11.823997551270278, "frac_reward_zero_std": 0.21250000670552255, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 1978518033.0, "reward": 7.909956645965576, "reward_std": 0.25213478654623034, "rewards/combined_reward/mean": 7.909956407546997, "rewards/combined_reward/std": 2.549025070667267, "step": 19320 }, { "clip_ratio/high_max": 0.0022621997064561582, "clip_ratio/high_mean": 0.0012350657906305665, "clip_ratio/low_mean": 0.0008435601949410436, "clip_ratio/low_min": 0.00021445701228609929, "clip_ratio/region_mean": 0.0020786259789019822, "completions/clipped_ratio": 0.0, "completions/max_length": 590.4, "completions/max_terminated_length": 590.4, "completions/mean_length": 201.29037017822264, "completions/mean_terminated_length": 201.29037017822264, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.03327204454690218, "epoch": 11.842363024181205, "frac_reward_zero_std": 0.20416667088866233, "grad_norm": 0.0267333984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 1981591148.0, "reward": 8.19636173248291, "reward_std": 0.3693172991275787, "rewards/combined_reward/mean": 8.196361303329468, "rewards/combined_reward/std": 2.0914090514183044, "step": 19350 }, { "clip_ratio/high_max": 0.0018044260917425465, "clip_ratio/high_mean": 0.0010395699466850298, "clip_ratio/low_mean": 0.0007448521480303801, "clip_ratio/low_min": 0.00024416965340302946, "clip_ratio/region_mean": 0.0017844220967769312, "completions/clipped_ratio": 0.0, "completions/max_length": 572.7, "completions/max_terminated_length": 572.7, "completions/mean_length": 205.10208892822266, "completions/mean_terminated_length": 205.10208892822266, "completions/min_length": 47.6, "completions/min_terminated_length": 47.6, "entropy": 0.0275869590540727, "epoch": 11.860728497092133, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.007110595703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 1984678900.0, "reward": 8.4438232421875, "reward_std": 0.2905083015561104, "rewards/combined_reward/mean": 8.443822765350342, "rewards/combined_reward/std": 1.7967964291572571, "step": 19380 }, { "clip_ratio/high_max": 0.0019258348445873707, "clip_ratio/high_mean": 0.001120948029711144, "clip_ratio/low_mean": 0.0007005966498885148, "clip_ratio/low_min": 0.00020605590943887365, "clip_ratio/region_mean": 0.0018215446781444675, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 209.95052642822264, "completions/mean_terminated_length": 209.95052642822264, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.030186481789375344, "epoch": 11.87909397000306, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 1987785270.0, "reward": 8.20148949623108, "reward_std": 0.30585474520921707, "rewards/combined_reward/mean": 8.201489210128784, "rewards/combined_reward/std": 2.125653553009033, "step": 19410 }, { "clip_ratio/high_max": 0.0017931718859472312, "clip_ratio/high_mean": 0.0010548486843617866, "clip_ratio/low_mean": 0.0007153382984445973, "clip_ratio/low_min": 0.0002132728213230924, "clip_ratio/region_mean": 0.0017701870002686823, "completions/clipped_ratio": 0.0, "completions/max_length": 594.8, "completions/max_terminated_length": 594.8, "completions/mean_length": 213.87838897705078, "completions/mean_terminated_length": 213.87838897705078, "completions/min_length": 50.4, "completions/min_terminated_length": 50.4, "entropy": 0.026817961192379395, "epoch": 11.897459442913988, "frac_reward_zero_std": 0.1625000026077032, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 1990906723.0, "reward": 8.084686517715454, "reward_std": 0.3214359626173973, "rewards/combined_reward/mean": 8.084686088562012, "rewards/combined_reward/std": 2.1393257856369017, "step": 19440 }, { "clip_ratio/high_max": 0.002086048118265656, "clip_ratio/high_mean": 0.001181069055989307, "clip_ratio/low_mean": 0.0008162142815611636, "clip_ratio/low_min": 0.00019628317216605258, "clip_ratio/region_mean": 0.0019972833251813427, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 191.64583587646484, "completions/mean_terminated_length": 191.64583587646484, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.031412884670620166, "epoch": 11.915824915824915, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.005950927734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 1993942803.0, "reward": 8.093414783477783, "reward_std": 0.34244067370891573, "rewards/combined_reward/mean": 8.093414688110352, "rewards/combined_reward/std": 2.0310988426208496, "step": 19470 }, { "clip_ratio/high_max": 0.0017432642979353355, "clip_ratio/high_mean": 0.0010253712908403637, "clip_ratio/low_mean": 0.0007646482773755755, "clip_ratio/low_min": 0.00024300522951913687, "clip_ratio/region_mean": 0.0017900195666394818, "completions/clipped_ratio": 0.0, "completions/max_length": 615.9, "completions/max_terminated_length": 615.9, "completions/mean_length": 206.25391235351563, "completions/mean_terminated_length": 206.25391235351563, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.029783710511401294, "epoch": 11.934190388735843, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.0038604736328125, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 1997034978.0, "reward": 8.150094032287598, "reward_std": 0.2306945726275444, "rewards/combined_reward/mean": 8.150093698501587, "rewards/combined_reward/std": 2.304136610031128, "step": 19500 }, { "clip_ratio/high_max": 0.0018765939263782154, "clip_ratio/high_mean": 0.0010468048004743955, "clip_ratio/low_mean": 0.0008638328616749883, "clip_ratio/low_min": 0.00023500000703885842, "clip_ratio/region_mean": 0.0019106376518417771, "completions/clipped_ratio": 0.0, "completions/max_length": 603.1, "completions/max_terminated_length": 603.1, "completions/mean_length": 205.58724517822264, "completions/mean_terminated_length": 205.58724517822264, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028182236912349858, "epoch": 11.95255586164677, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2000124593.0, "reward": 7.88954610824585, "reward_std": 0.28803258910775187, "rewards/combined_reward/mean": 7.889545869827271, "rewards/combined_reward/std": 2.343931758403778, "step": 19530 }, { "clip_ratio/high_max": 0.0015392226002101477, "clip_ratio/high_mean": 0.000894616097624142, "clip_ratio/low_mean": 0.0007533547786200263, "clip_ratio/low_min": 0.0002453941859130282, "clip_ratio/region_mean": 0.0016479708753953067, "completions/clipped_ratio": 0.0, "completions/max_length": 568.5, "completions/max_terminated_length": 568.5, "completions/mean_length": 193.43333892822267, "completions/mean_terminated_length": 193.43333892822267, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.02568873066144685, "epoch": 11.970921334557698, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.0145263671875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2003167537.0, "reward": 8.072953987121583, "reward_std": 0.27384268641471865, "rewards/combined_reward/mean": 8.072953701019287, "rewards/combined_reward/std": 2.351943290233612, "step": 19560 }, { "clip_ratio/high_max": 0.0015539992518218545, "clip_ratio/high_mean": 0.0009404028423887212, "clip_ratio/low_mean": 0.0007586441875901073, "clip_ratio/low_min": 0.0002452552895798969, "clip_ratio/region_mean": 0.0016990470302213603, "completions/clipped_ratio": 0.0, "completions/max_length": 632.8, "completions/max_terminated_length": 632.8, "completions/mean_length": 205.6901092529297, "completions/mean_terminated_length": 205.6901092529297, "completions/min_length": 46.7, "completions/min_terminated_length": 46.7, "entropy": 0.0244112824400266, "epoch": 11.989286807468625, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2006257547.0, "reward": 8.084255123138428, "reward_std": 0.3213101699948311, "rewards/combined_reward/mean": 8.084254789352418, "rewards/combined_reward/std": 2.2190094113349916, "step": 19590 }, { "clip_ratio/high_max": 0.002142867953926859, "clip_ratio/high_mean": 0.0012560367415330876, "clip_ratio/low_mean": 0.0008389461730925565, "clip_ratio/low_min": 0.0002764186842870883, "clip_ratio/region_mean": 0.0020949829228558256, "completions/clipped_ratio": 0.0, "completions/max_length": 509.6, "completions/max_terminated_length": 509.6, "completions/mean_length": 203.91797332763673, "completions/mean_terminated_length": 203.91797332763673, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.028429542362812112, "epoch": 12.007958371594736, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 2009340752.0, "reward": 8.129360580444336, "reward_std": 0.362348672747612, "rewards/combined_reward/mean": 8.12936053276062, "rewards/combined_reward/std": 2.2514520645141602, "step": 19620 }, { "clip_ratio/high_max": 0.002624692623066949, "clip_ratio/high_mean": 0.0015052196983257697, "clip_ratio/low_mean": 0.0010923545032710535, "clip_ratio/low_min": 0.0003688182907353621, "clip_ratio/region_mean": 0.0025975742042646742, "completions/clipped_ratio": 0.0, "completions/max_length": 475.2, "completions/max_terminated_length": 475.2, "completions/mean_length": 195.78906707763673, "completions/mean_terminated_length": 195.78906707763673, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028983069692427914, "epoch": 12.026323844505663, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2012392742.0, "reward": 8.089617109298706, "reward_std": 0.3163066178560257, "rewards/combined_reward/mean": 8.089616680145264, "rewards/combined_reward/std": 2.1939609050750732, "step": 19650 }, { "clip_ratio/high_max": 0.002981378634770711, "clip_ratio/high_mean": 0.0017143111472250894, "clip_ratio/low_mean": 0.0010321121631932328, "clip_ratio/low_min": 0.0002986841105060497, "clip_ratio/region_mean": 0.0027464232819814547, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 209.06146392822265, "completions/mean_terminated_length": 209.06146392822265, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.03361575352028012, "epoch": 12.04468931741659, "frac_reward_zero_std": 0.17083333991467953, "grad_norm": 0.0042724609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2015495698.0, "reward": 7.970173978805542, "reward_std": 0.3712351739406586, "rewards/combined_reward/mean": 7.970173740386963, "rewards/combined_reward/std": 2.3171762585639955, "step": 19680 }, { "clip_ratio/high_max": 0.0020468460184929426, "clip_ratio/high_mean": 0.0012319390916673, "clip_ratio/low_mean": 0.0008993734795997928, "clip_ratio/low_min": 0.00024714634879880273, "clip_ratio/region_mean": 0.0021313125595042947, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 201.39818267822267, "completions/mean_terminated_length": 201.39818267822267, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.025339359727998573, "epoch": 12.063054790327518, "frac_reward_zero_std": 0.1375000048428774, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2018569227.0, "reward": 8.296552658081055, "reward_std": 0.27460883408784864, "rewards/combined_reward/mean": 8.296552228927613, "rewards/combined_reward/std": 2.09898955821991, "step": 19710 }, { "clip_ratio/high_max": 0.002501676625009471, "clip_ratio/high_mean": 0.0013140735132765257, "clip_ratio/low_mean": 0.0009473788561687494, "clip_ratio/low_min": 0.00022869506186301198, "clip_ratio/region_mean": 0.0022614523388862534, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 196.64167022705078, "completions/mean_terminated_length": 196.64167022705078, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.02957440197157363, "epoch": 12.081420263238446, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2021624491.0, "reward": 8.252772521972656, "reward_std": 0.2854266718029976, "rewards/combined_reward/mean": 8.252772283554076, "rewards/combined_reward/std": 1.860984206199646, "step": 19740 }, { "clip_ratio/high_max": 0.002127303255353278, "clip_ratio/high_mean": 0.0012197643451751599, "clip_ratio/low_mean": 0.00086925759484681, "clip_ratio/low_min": 0.00028516538441181183, "clip_ratio/region_mean": 0.0020890219345650013, "completions/clipped_ratio": 0.0, "completions/max_length": 639.9, "completions/max_terminated_length": 639.9, "completions/mean_length": 207.37240142822264, "completions/mean_terminated_length": 207.37240142822264, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.030093465466052294, "epoch": 12.099785736149373, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2024720961.0, "reward": 8.024740314483642, "reward_std": 0.3240070253610611, "rewards/combined_reward/mean": 8.024740171432494, "rewards/combined_reward/std": 2.1791940212249754, "step": 19770 }, { "clip_ratio/high_max": 0.0023372191057812112, "clip_ratio/high_mean": 0.0013251795280666556, "clip_ratio/low_mean": 0.0010521640842246901, "clip_ratio/low_min": 0.0004111653555204005, "clip_ratio/region_mean": 0.002377343608532101, "completions/clipped_ratio": 0.0, "completions/max_length": 641.5, "completions/max_terminated_length": 641.5, "completions/mean_length": 216.88776397705078, "completions/mean_terminated_length": 216.88776397705078, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.030728524442141254, "epoch": 12.1181512090603, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2027853970.0, "reward": 7.935199737548828, "reward_std": 0.2904248028993607, "rewards/combined_reward/mean": 7.935199499130249, "rewards/combined_reward/std": 2.4838048338890077, "step": 19800 }, { "clip_ratio/high_max": 0.0024073224657816658, "clip_ratio/high_mean": 0.0013682650028689143, "clip_ratio/low_mean": 0.0009730218015950716, "clip_ratio/low_min": 0.0002482104086084291, "clip_ratio/region_mean": 0.00234128680434272, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 186.05912017822266, "completions/mean_terminated_length": 186.05912017822266, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02973484150910129, "epoch": 12.136516681971228, "frac_reward_zero_std": 0.20000000707805157, "grad_norm": 0.0038299560546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2030868597.0, "reward": 8.002217006683349, "reward_std": 0.22638382613658906, "rewards/combined_reward/mean": 8.002216815948486, "rewards/combined_reward/std": 2.315653645992279, "step": 19830 }, { "clip_ratio/high_max": 0.0021872147796481536, "clip_ratio/high_mean": 0.0012180440253966177, "clip_ratio/low_mean": 0.0009844592185496972, "clip_ratio/low_min": 0.00023692961476626807, "clip_ratio/region_mean": 0.0022025032529199963, "completions/clipped_ratio": 0.0, "completions/max_length": 526.6, "completions/max_terminated_length": 526.6, "completions/mean_length": 201.96693267822266, "completions/mean_terminated_length": 201.96693267822266, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.02824258526476721, "epoch": 12.154882154882156, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.004486083984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2033944310.0, "reward": 8.104472589492797, "reward_std": 0.33272538930177686, "rewards/combined_reward/mean": 8.104472208023072, "rewards/combined_reward/std": 2.0913413643836973, "step": 19860 }, { "clip_ratio/high_max": 0.0021572911685022217, "clip_ratio/high_mean": 0.001215547785856567, "clip_ratio/low_mean": 0.0008558612452664722, "clip_ratio/low_min": 0.0001943616278974029, "clip_ratio/region_mean": 0.0020714090092951664, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 208.43021392822266, "completions/mean_terminated_length": 208.43021392822266, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.026834279710116486, "epoch": 12.173247627793081, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.005279541015625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2037044842.0, "reward": 7.990139961242676, "reward_std": 0.2927345111966133, "rewards/combined_reward/mean": 7.990139579772949, "rewards/combined_reward/std": 2.375309908390045, "step": 19890 }, { "clip_ratio/high_max": 0.0023142442795991277, "clip_ratio/high_mean": 0.0012627557114077111, "clip_ratio/low_mean": 0.0009268575568664043, "clip_ratio/low_min": 0.0002701854828046635, "clip_ratio/region_mean": 0.0021896132699718387, "completions/clipped_ratio": 0.0, "completions/max_length": 553.4, "completions/max_terminated_length": 553.4, "completions/mean_length": 196.30755767822265, "completions/mean_terminated_length": 196.30755767822265, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027185659793516, "epoch": 12.191613100704009, "frac_reward_zero_std": 0.20833334065973758, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2040098823.0, "reward": 8.094255495071412, "reward_std": 0.24846612066030502, "rewards/combined_reward/mean": 8.094255304336547, "rewards/combined_reward/std": 2.2794673442840576, "step": 19920 }, { "clip_ratio/high_max": 0.0023259287159210847, "clip_ratio/high_mean": 0.0013394698629175157, "clip_ratio/low_mean": 0.0008451337309452356, "clip_ratio/low_min": 0.00019306435545634788, "clip_ratio/region_mean": 0.0021846035922256607, "completions/clipped_ratio": 0.0, "completions/max_length": 587.6, "completions/max_terminated_length": 587.6, "completions/mean_length": 199.1210968017578, "completions/mean_terminated_length": 199.1210968017578, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.029091270423183838, "epoch": 12.209978573614936, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.007171630859375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2043163608.0, "reward": 8.165374994277954, "reward_std": 0.25680516064167025, "rewards/combined_reward/mean": 8.165374374389648, "rewards/combined_reward/std": 2.19407719373703, "step": 19950 }, { "clip_ratio/high_max": 0.0024595083901658653, "clip_ratio/high_mean": 0.001374804933705794, "clip_ratio/low_mean": 0.0010296804895309228, "clip_ratio/low_min": 0.0003281660940653334, "clip_ratio/region_mean": 0.002404485403288466, "completions/clipped_ratio": 0.0, "completions/max_length": 608.4, "completions/max_terminated_length": 608.4, "completions/mean_length": 199.53125457763673, "completions/mean_terminated_length": 199.53125457763673, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03257003913943966, "epoch": 12.228344046525864, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2046229968.0, "reward": 8.065193462371827, "reward_std": 0.2660456508398056, "rewards/combined_reward/mean": 8.065193223953248, "rewards/combined_reward/std": 2.26620010137558, "step": 19980 }, { "clip_ratio/high_max": 0.0021383747453607308, "clip_ratio/high_mean": 0.0012312731535833639, "clip_ratio/low_mean": 0.0009330739201080481, "clip_ratio/low_min": 0.00028291513493362196, "clip_ratio/region_mean": 0.002164347045860874, "completions/clipped_ratio": 0.0, "completions/max_length": 593.9, "completions/max_terminated_length": 593.9, "completions/mean_length": 198.24844360351562, "completions/mean_terminated_length": 198.24844360351562, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.028172985355680188, "epoch": 12.246709519436791, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.006134033203125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2049291402.0, "reward": 8.140707540512086, "reward_std": 0.3048986494541168, "rewards/combined_reward/mean": 8.140707445144653, "rewards/combined_reward/std": 2.107535243034363, "step": 20010 }, { "clip_ratio/high_max": 0.002137424180788609, "clip_ratio/high_mean": 0.0012145280659751733, "clip_ratio/low_mean": 0.0009703048559458694, "clip_ratio/low_min": 0.00028166670041779676, "clip_ratio/region_mean": 0.002184832930894724, "completions/clipped_ratio": 0.0, "completions/max_length": 609.4, "completions/max_terminated_length": 609.4, "completions/mean_length": 198.95521392822266, "completions/mean_terminated_length": 198.95521392822266, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.02768351879591743, "epoch": 12.265074992347719, "frac_reward_zero_std": 0.1750000063329935, "grad_norm": 0.007537841796875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2052355550.0, "reward": 8.118322515487671, "reward_std": 0.24155261740088463, "rewards/combined_reward/mean": 8.118322134017944, "rewards/combined_reward/std": 2.2018903613090517, "step": 20040 }, { "clip_ratio/high_max": 0.0022754635487217457, "clip_ratio/high_mean": 0.0012417094800184713, "clip_ratio/low_mean": 0.0010005002096780422, "clip_ratio/low_min": 0.0002460040945152286, "clip_ratio/region_mean": 0.0022422096895752474, "completions/clipped_ratio": 0.0, "completions/max_length": 637.5, "completions/max_terminated_length": 637.5, "completions/mean_length": 192.8937561035156, "completions/mean_terminated_length": 192.8937561035156, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029686537059023976, "epoch": 12.283440465258646, "frac_reward_zero_std": 0.19583333916962148, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2055396422.0, "reward": 7.948816394805908, "reward_std": 0.296563733369112, "rewards/combined_reward/mean": 7.948816633224487, "rewards/combined_reward/std": 2.37920595407486, "step": 20070 }, { "clip_ratio/high_max": 0.0024208636374775474, "clip_ratio/high_mean": 0.0013454374031425687, "clip_ratio/low_mean": 0.0009797568962918982, "clip_ratio/low_min": 0.0002952296602112862, "clip_ratio/region_mean": 0.002325194283184828, "completions/clipped_ratio": 0.0, "completions/max_length": 576.3, "completions/max_terminated_length": 576.3, "completions/mean_length": 199.75339050292968, "completions/mean_terminated_length": 199.75339050292968, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.026760501647368075, "epoch": 12.301805938169574, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.00634765625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2058463635.0, "reward": 8.014816665649414, "reward_std": 0.35984216779470446, "rewards/combined_reward/mean": 8.014816284179688, "rewards/combined_reward/std": 2.2337859988212587, "step": 20100 }, { "clip_ratio/high_max": 0.0025554597208004757, "clip_ratio/high_mean": 0.0015049066047746844, "clip_ratio/low_mean": 0.0010423377231442525, "clip_ratio/low_min": 0.00039182795905314075, "clip_ratio/region_mean": 0.0025472443288890646, "completions/clipped_ratio": 0.0, "completions/max_length": 584.3, "completions/max_terminated_length": 584.3, "completions/mean_length": 202.06094207763672, "completions/mean_terminated_length": 202.06094207763672, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03052614174472789, "epoch": 12.320171411080501, "frac_reward_zero_std": 0.11250000298023224, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2061539709.0, "reward": 8.31196722984314, "reward_std": 0.38833033442497256, "rewards/combined_reward/mean": 8.311966896057129, "rewards/combined_reward/std": 1.92804257273674, "step": 20130 }, { "clip_ratio/high_max": 0.0025231223303611235, "clip_ratio/high_mean": 0.0015015705229719364, "clip_ratio/low_mean": 0.0010540783857625987, "clip_ratio/low_min": 0.0002815719519276172, "clip_ratio/region_mean": 0.002555648886482231, "completions/clipped_ratio": 0.0, "completions/max_length": 550.5, "completions/max_terminated_length": 550.5, "completions/mean_length": 204.70130615234376, "completions/mean_terminated_length": 204.70130615234376, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.032372916117310524, "epoch": 12.338536883991429, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2064625922.0, "reward": 7.962127780914306, "reward_std": 0.32926389276981355, "rewards/combined_reward/mean": 7.962127685546875, "rewards/combined_reward/std": 2.54761917591095, "step": 20160 }, { "clip_ratio/high_max": 0.002455725425291651, "clip_ratio/high_mean": 0.0014289805035029228, "clip_ratio/low_mean": 0.0008771041183839164, "clip_ratio/low_min": 0.0002714823730154118, "clip_ratio/region_mean": 0.0023060846321944456, "completions/clipped_ratio": 0.0, "completions/max_length": 629.5, "completions/max_terminated_length": 629.5, "completions/mean_length": 200.96589050292968, "completions/mean_terminated_length": 200.96589050292968, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029567831909904876, "epoch": 12.356902356902356, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2067697791.0, "reward": 8.17459683418274, "reward_std": 0.3061296835541725, "rewards/combined_reward/mean": 8.174596548080444, "rewards/combined_reward/std": 2.182126152515411, "step": 20190 }, { "clip_ratio/high_max": 0.0023245960701994287, "clip_ratio/high_mean": 0.0013114751264462635, "clip_ratio/low_mean": 0.0009895423173778303, "clip_ratio/low_min": 0.0002667562677136933, "clip_ratio/region_mean": 0.002301017431212434, "completions/clipped_ratio": 0.0, "completions/max_length": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 206.38385925292968, "completions/mean_terminated_length": 206.38385925292968, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02835286979873975, "epoch": 12.375267829813284, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2070790465.0, "reward": 8.082223987579345, "reward_std": 0.30682533979415894, "rewards/combined_reward/mean": 8.082223463058472, "rewards/combined_reward/std": 2.258944070339203, "step": 20220 }, { "clip_ratio/high_max": 0.0026975290895885945, "clip_ratio/high_mean": 0.0015592734658033199, "clip_ratio/low_mean": 0.0011451671094012757, "clip_ratio/low_min": 0.0003922278886117662, "clip_ratio/region_mean": 0.0027044405578635633, "completions/clipped_ratio": 0.0, "completions/max_length": 599.3, "completions/max_terminated_length": 599.3, "completions/mean_length": 202.31771240234374, "completions/mean_terminated_length": 202.31771240234374, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.032436927671854694, "epoch": 12.393633302724211, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.031982421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2073867525.0, "reward": 7.758608055114746, "reward_std": 0.37053926587104796, "rewards/combined_reward/mean": 7.758607721328735, "rewards/combined_reward/std": 2.446167993545532, "step": 20250 }, { "clip_ratio/high_max": 0.002858838700437142, "clip_ratio/high_mean": 0.001509337355309981, "clip_ratio/low_mean": 0.0009275445017313662, "clip_ratio/low_min": 0.0002566434459974213, "clip_ratio/region_mean": 0.0024368818344858785, "completions/clipped_ratio": 0.0, "completions/max_length": 517.8, "completions/max_terminated_length": 517.8, "completions/mean_length": 195.37109832763673, "completions/mean_terminated_length": 195.37109832763673, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.02944937056551377, "epoch": 12.411998775635139, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.003997802734375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2076917910.0, "reward": 7.914537668228149, "reward_std": 0.3243178606033325, "rewards/combined_reward/mean": 7.914537334442139, "rewards/combined_reward/std": 2.321390676498413, "step": 20280 }, { "clip_ratio/high_max": 0.0025378438830860736, "clip_ratio/high_mean": 0.001403628516709432, "clip_ratio/low_mean": 0.000853240852302406, "clip_ratio/low_min": 0.0002246978472006352, "clip_ratio/region_mean": 0.002256869378349317, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 206.43932495117187, "completions/mean_terminated_length": 206.43932495117187, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.028094747662544252, "epoch": 12.430364248546066, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2080010797.0, "reward": 7.8714994430542, "reward_std": 0.32425723820924757, "rewards/combined_reward/mean": 7.871499109268188, "rewards/combined_reward/std": 2.576585364341736, "step": 20310 }, { "clip_ratio/high_max": 0.00240056308102794, "clip_ratio/high_mean": 0.0014258537902302729, "clip_ratio/low_mean": 0.0009038348275983783, "clip_ratio/low_min": 0.00028047040056359646, "clip_ratio/region_mean": 0.002329688610431428, "completions/clipped_ratio": 0.0, "completions/max_length": 611.5, "completions/max_terminated_length": 611.5, "completions/mean_length": 206.25234985351562, "completions/mean_terminated_length": 206.25234985351562, "completions/min_length": 47.0, "completions/min_terminated_length": 47.0, "entropy": 0.030407162251261374, "epoch": 12.448729721456994, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 2083102966.0, "reward": 8.258435344696045, "reward_std": 0.2924749568104744, "rewards/combined_reward/mean": 8.258435153961182, "rewards/combined_reward/std": 2.2697630524635315, "step": 20340 }, { "clip_ratio/high_max": 0.0030697491437119123, "clip_ratio/high_mean": 0.0017986586955278958, "clip_ratio/low_mean": 0.0009931076490223253, "clip_ratio/low_min": 0.00028000374125743593, "clip_ratio/region_mean": 0.002791766319811965, "completions/clipped_ratio": 0.0, "completions/max_length": 534.8, "completions/max_terminated_length": 534.8, "completions/mean_length": 198.3156295776367, "completions/mean_terminated_length": 198.3156295776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03491993703258534, "epoch": 12.467095194367921, "frac_reward_zero_std": 0.2041666705161333, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2086164658.0, "reward": 7.932775783538818, "reward_std": 0.4006653562188148, "rewards/combined_reward/mean": 7.932775354385376, "rewards/combined_reward/std": 2.2912067413330077, "step": 20370 }, { "clip_ratio/high_max": 0.0026794205832023485, "clip_ratio/high_mean": 0.0014881620547384955, "clip_ratio/low_mean": 0.0009726476966534392, "clip_ratio/low_min": 0.00030194864035972083, "clip_ratio/region_mean": 0.0024608097548480144, "completions/clipped_ratio": 0.0, "completions/max_length": 512.1, "completions/max_terminated_length": 512.1, "completions/mean_length": 191.5304718017578, "completions/mean_terminated_length": 191.5304718017578, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02833116172502438, "epoch": 12.485460667278849, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2089200295.0, "reward": 7.969076299667359, "reward_std": 0.40542994141578675, "rewards/combined_reward/mean": 7.969076013565063, "rewards/combined_reward/std": 2.308334243297577, "step": 20400 }, { "clip_ratio/high_max": 0.002294314963607273, "clip_ratio/high_mean": 0.0012690124574874062, "clip_ratio/low_mean": 0.0007782986432478841, "clip_ratio/low_min": 0.0002202826988650486, "clip_ratio/region_mean": 0.0020473111003714924, "completions/clipped_ratio": 0.0, "completions/max_length": 562.6, "completions/max_terminated_length": 562.6, "completions/mean_length": 198.94193267822266, "completions/mean_terminated_length": 198.94193267822266, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03275104363759359, "epoch": 12.503213957759412, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.0291748046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2092264392.0, "reward": 7.9003266334533695, "reward_std": 0.35195759534835813, "rewards/combined_reward/mean": 7.900326490402222, "rewards/combined_reward/std": 2.4036492466926576, "step": 20430 }, { "clip_ratio/high_max": 0.0017722349815206446, "clip_ratio/high_mean": 0.0009763539015693823, "clip_ratio/low_mean": 0.0007209130982422115, "clip_ratio/low_min": 0.00019464642779591182, "clip_ratio/region_mean": 0.0016972669909591787, "completions/clipped_ratio": 0.0, "completions/max_length": 565.4, "completions/max_terminated_length": 565.4, "completions/mean_length": 194.17552642822267, "completions/mean_terminated_length": 194.17552642822267, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02710789969811837, "epoch": 12.52157943067034, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2095310186.0, "reward": 7.885613536834716, "reward_std": 0.2989142417907715, "rewards/combined_reward/mean": 7.885613346099854, "rewards/combined_reward/std": 2.5544071555137635, "step": 20460 }, { "clip_ratio/high_max": 0.0018967624637298286, "clip_ratio/high_mean": 0.001062313114622763, "clip_ratio/low_mean": 0.0007804577970091486, "clip_ratio/low_min": 0.00020758007206798842, "clip_ratio/region_mean": 0.001842770928730412, "completions/clipped_ratio": 0.0, "completions/max_length": 583.9, "completions/max_terminated_length": 583.9, "completions/mean_length": 197.40209045410157, "completions/mean_terminated_length": 197.40209045410157, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.02866557027834157, "epoch": 12.539944903581267, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.00933837890625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2098368370.0, "reward": 8.040665531158448, "reward_std": 0.2929134860634804, "rewards/combined_reward/mean": 8.040665197372437, "rewards/combined_reward/std": 2.3101009011268614, "step": 20490 }, { "clip_ratio/high_max": 0.0016779497003881262, "clip_ratio/high_mean": 0.000989594370306198, "clip_ratio/low_mean": 0.0007548186948649042, "clip_ratio/low_min": 0.00021393500768075076, "clip_ratio/region_mean": 0.0017444130741447832, "completions/clipped_ratio": 0.0, "completions/max_length": 582.9, "completions/max_terminated_length": 582.9, "completions/mean_length": 197.5773483276367, "completions/mean_terminated_length": 197.5773483276367, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02523156232200563, "epoch": 12.558310376492194, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.005218505859375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2101427227.0, "reward": 8.364489364624024, "reward_std": 0.2092433214187622, "rewards/combined_reward/mean": 8.364489030838012, "rewards/combined_reward/std": 1.8788600385189056, "step": 20520 }, { "clip_ratio/high_max": 0.002305852407638061, "clip_ratio/high_mean": 0.0013291199726154445, "clip_ratio/low_mean": 0.0008782089923139816, "clip_ratio/low_min": 0.00021470351760702517, "clip_ratio/region_mean": 0.0022073289457087714, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 208.52969207763672, "completions/mean_terminated_length": 208.52969207763672, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 0.03459498795370261, "epoch": 12.576675849403122, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2104528141.0, "reward": 7.673199129104614, "reward_std": 0.3425112426280975, "rewards/combined_reward/mean": 7.6731987476348875, "rewards/combined_reward/std": 2.5740591049194337, "step": 20550 }, { "clip_ratio/high_max": 0.0015374682673912805, "clip_ratio/high_mean": 0.0008746331085906907, "clip_ratio/low_mean": 0.000784862327054725, "clip_ratio/low_min": 0.00023821344439056702, "clip_ratio/region_mean": 0.0016594954237613517, "completions/clipped_ratio": 0.0, "completions/max_length": 559.1, "completions/max_terminated_length": 559.1, "completions/mean_length": 216.8205764770508, "completions/mean_terminated_length": 216.8205764770508, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.026089670214181144, "epoch": 12.59504132231405, "frac_reward_zero_std": 0.15833333916962147, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2107660892.0, "reward": 8.081426906585694, "reward_std": 0.33779042214155197, "rewards/combined_reward/mean": 8.081426429748536, "rewards/combined_reward/std": 2.0353887557983397, "step": 20580 }, { "clip_ratio/high_max": 0.0019118608911715758, "clip_ratio/high_mean": 0.001030402577574326, "clip_ratio/low_mean": 0.0008237866179115372, "clip_ratio/low_min": 0.0002626936290956413, "clip_ratio/region_mean": 0.0018541891675946923, "completions/clipped_ratio": 0.0, "completions/max_length": 619.2, "completions/max_terminated_length": 619.2, "completions/mean_length": 202.37162170410156, "completions/mean_terminated_length": 202.37162170410156, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.028848934111495812, "epoch": 12.613406795224977, "frac_reward_zero_std": 0.14583333544433116, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2110738159.0, "reward": 8.036549472808838, "reward_std": 0.33971632421016695, "rewards/combined_reward/mean": 8.036549043655395, "rewards/combined_reward/std": 2.1682691097259523, "step": 20610 }, { "clip_ratio/high_max": 0.0017089208631659858, "clip_ratio/high_mean": 0.0009020593446621205, "clip_ratio/low_mean": 0.000764284422621131, "clip_ratio/low_min": 0.00028093732180423103, "clip_ratio/region_mean": 0.0016663437408472722, "completions/clipped_ratio": 0.0, "completions/max_length": 637.4, "completions/max_terminated_length": 637.4, "completions/mean_length": 204.05833892822267, "completions/mean_terminated_length": 204.05833892822267, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.028319642630716166, "epoch": 12.631772268135904, "frac_reward_zero_std": 0.12500000409781933, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2113821903.0, "reward": 8.290156936645507, "reward_std": 0.2624910339713097, "rewards/combined_reward/mean": 8.290156507492066, "rewards/combined_reward/std": 2.101172482967377, "step": 20640 }, { "clip_ratio/high_max": 0.002042563069699099, "clip_ratio/high_mean": 0.0012039921010000398, "clip_ratio/low_mean": 0.0007656252205682297, "clip_ratio/low_min": 0.00025452319387113674, "clip_ratio/region_mean": 0.0019696173369690467, "completions/clipped_ratio": 0.0, "completions/max_length": 544.5, "completions/max_terminated_length": 544.5, "completions/mean_length": 208.63724517822266, "completions/mean_terminated_length": 208.63724517822266, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03132418125557403, "epoch": 12.650137741046832, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2116923230.0, "reward": 8.00691556930542, "reward_std": 0.33817571550607683, "rewards/combined_reward/mean": 8.006915473937989, "rewards/combined_reward/std": 2.21192409992218, "step": 20670 }, { "clip_ratio/high_max": 0.0017421442923174861, "clip_ratio/high_mean": 0.0009984370693321882, "clip_ratio/low_mean": 0.0007203905336306585, "clip_ratio/low_min": 0.0002131645756890066, "clip_ratio/region_mean": 0.0017188276042967724, "completions/clipped_ratio": 0.0, "completions/max_length": 522.3, "completions/max_terminated_length": 522.3, "completions/mean_length": 203.89740142822265, "completions/mean_terminated_length": 203.89740142822265, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028191454987972975, "epoch": 12.66850321395776, "frac_reward_zero_std": 0.1791666742414236, "grad_norm": 0.038818359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2120006356.0, "reward": 7.848360013961792, "reward_std": 0.36979087591171267, "rewards/combined_reward/mean": 7.848359680175781, "rewards/combined_reward/std": 2.3119063854217528, "step": 20700 }, { "clip_ratio/high_max": 0.0018077241348995207, "clip_ratio/high_mean": 0.0010162492043794677, "clip_ratio/low_mean": 0.0007280758860967277, "clip_ratio/low_min": 0.00027150287084320247, "clip_ratio/region_mean": 0.0017443250801685887, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 195.5265686035156, "completions/mean_terminated_length": 195.5265686035156, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.026697966487457356, "epoch": 12.686868686868687, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2123057338.0, "reward": 8.141280174255371, "reward_std": 0.3021052241325378, "rewards/combined_reward/mean": 8.141280031204223, "rewards/combined_reward/std": 2.1679829001426696, "step": 20730 }, { "clip_ratio/high_max": 0.0021033248400878315, "clip_ratio/high_mean": 0.0010635672912030713, "clip_ratio/low_mean": 0.0007215465769453052, "clip_ratio/low_min": 0.0001378779406271254, "clip_ratio/region_mean": 0.0017851138763944618, "completions/clipped_ratio": 0.0, "completions/max_length": 534.4, "completions/max_terminated_length": 534.4, "completions/mean_length": 198.9158920288086, "completions/mean_terminated_length": 198.9158920288086, "completions/min_length": 47.6, "completions/min_terminated_length": 47.6, "entropy": 0.027950138226151465, "epoch": 12.705234159779614, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.005340576171875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2126121335.0, "reward": 8.228067016601562, "reward_std": 0.23672997057437897, "rewards/combined_reward/mean": 8.228066635131835, "rewards/combined_reward/std": 1.9741195380687713, "step": 20760 }, { "clip_ratio/high_max": 0.0021998710217303596, "clip_ratio/high_mean": 0.0012263184615828018, "clip_ratio/low_mean": 0.0009251089030537211, "clip_ratio/low_min": 0.0002533784985037831, "clip_ratio/region_mean": 0.0021514273578456295, "completions/clipped_ratio": 0.0, "completions/max_length": 595.4, "completions/max_terminated_length": 595.4, "completions/mean_length": 220.85937957763673, "completions/mean_terminated_length": 220.85937957763673, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.030583720305003226, "epoch": 12.723599632690542, "frac_reward_zero_std": 0.12083333618938923, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2129269595.0, "reward": 8.195319509506225, "reward_std": 0.3249676004052162, "rewards/combined_reward/mean": 8.195319223403931, "rewards/combined_reward/std": 2.10734224319458, "step": 20790 }, { "clip_ratio/high_max": 0.001905200141967119, "clip_ratio/high_mean": 0.0011086266923181634, "clip_ratio/low_mean": 0.0007077984456069923, "clip_ratio/low_min": 0.0001787041774756896, "clip_ratio/region_mean": 0.0018164251499304858, "completions/clipped_ratio": 0.0, "completions/max_length": 522.3, "completions/max_terminated_length": 522.3, "completions/mean_length": 204.13282012939453, "completions/mean_terminated_length": 204.13282012939453, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027049655125786862, "epoch": 12.74196510560147, "frac_reward_zero_std": 0.18333333767950535, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2132353625.0, "reward": 8.073338747024536, "reward_std": 0.3240445092320442, "rewards/combined_reward/mean": 8.07333846092224, "rewards/combined_reward/std": 2.3875523447990417, "step": 20820 }, { "clip_ratio/high_max": 0.001802079483847289, "clip_ratio/high_mean": 0.0011098540746994937, "clip_ratio/low_mean": 0.0006826887029698507, "clip_ratio/low_min": 0.00019122395387967118, "clip_ratio/region_mean": 0.0017925427741526316, "completions/clipped_ratio": 0.0, "completions/max_length": 603.7, "completions/max_terminated_length": 603.7, "completions/mean_length": 203.7252670288086, "completions/mean_terminated_length": 203.7252670288086, "completions/min_length": 50.3, "completions/min_terminated_length": 50.3, "entropy": 0.028924606360184648, "epoch": 12.760330578512397, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0247802734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2135436090.0, "reward": 8.110758209228516, "reward_std": 0.33215630501508714, "rewards/combined_reward/mean": 8.110758209228516, "rewards/combined_reward/std": 2.165544641017914, "step": 20850 }, { "clip_ratio/high_max": 0.00187838773514765, "clip_ratio/high_mean": 0.0010693586019139426, "clip_ratio/low_mean": 0.000678044533560751, "clip_ratio/low_min": 0.00016996661142911761, "clip_ratio/region_mean": 0.0017474031250458212, "completions/clipped_ratio": 0.0, "completions/max_length": 542.5, "completions/max_terminated_length": 542.5, "completions/mean_length": 213.10391082763672, "completions/mean_terminated_length": 213.10391082763672, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.027931642757418256, "epoch": 12.778696051423324, "frac_reward_zero_std": 0.18333333618938924, "grad_norm": 0.0250244140625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2138554569.0, "reward": 8.208896207809449, "reward_std": 0.29088445454835893, "rewards/combined_reward/mean": 8.208896160125732, "rewards/combined_reward/std": 2.163522458076477, "step": 20880 }, { "clip_ratio/high_max": 0.0021802213884560234, "clip_ratio/high_mean": 0.0011861131123926802, "clip_ratio/low_mean": 0.0007473883704127123, "clip_ratio/low_min": 0.00022662517755331162, "clip_ratio/region_mean": 0.001933501495902116, "completions/clipped_ratio": 0.0, "completions/max_length": 530.1, "completions/max_terminated_length": 530.1, "completions/mean_length": 202.4234420776367, "completions/mean_terminated_length": 202.4234420776367, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.03028503411139051, "epoch": 12.797061524334252, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2141632035.0, "reward": 8.091301965713502, "reward_std": 0.30614923536777494, "rewards/combined_reward/mean": 8.091301822662354, "rewards/combined_reward/std": 2.233031117916107, "step": 20910 }, { "clip_ratio/high_max": 0.002398616939414448, "clip_ratio/high_mean": 0.0013779412060102914, "clip_ratio/low_mean": 0.0008745141261897515, "clip_ratio/low_min": 0.00024659652141660143, "clip_ratio/region_mean": 0.002252455313767617, "completions/clipped_ratio": 0.0, "completions/max_length": 617.7, "completions/max_terminated_length": 617.7, "completions/mean_length": 195.7143295288086, "completions/mean_terminated_length": 195.7143295288086, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03147904205446442, "epoch": 12.81542699724518, "frac_reward_zero_std": 0.20416667275130748, "grad_norm": 0.00628662109375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2144683738.0, "reward": 7.905479669570923, "reward_std": 0.26939839869737625, "rewards/combined_reward/mean": 7.9054793357849125, "rewards/combined_reward/std": 2.3450502514839173, "step": 20940 }, { "clip_ratio/high_max": 0.002655870239929451, "clip_ratio/high_mean": 0.0013268252754642163, "clip_ratio/low_mean": 0.0008725525953195756, "clip_ratio/low_min": 0.0002779309994366486, "clip_ratio/region_mean": 0.002199377860718717, "completions/clipped_ratio": 0.0, "completions/max_length": 600.4, "completions/max_terminated_length": 600.4, "completions/mean_length": 208.2955764770508, "completions/mean_terminated_length": 208.2955764770508, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.03050465345537911, "epoch": 12.833792470156107, "frac_reward_zero_std": 0.1833333358168602, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2147783753.0, "reward": 8.247215938568115, "reward_std": 0.32183659821748734, "rewards/combined_reward/mean": 8.247215747833252, "rewards/combined_reward/std": 2.1386165022850037, "step": 20970 }, { "clip_ratio/high_max": 0.0020409339704201557, "clip_ratio/high_mean": 0.0011506328319228487, "clip_ratio/low_mean": 0.0007395931440745092, "clip_ratio/low_min": 0.00019645030867347182, "clip_ratio/region_mean": 0.0018902259718743153, "completions/clipped_ratio": 0.0, "completions/max_length": 654.1, "completions/max_terminated_length": 654.1, "completions/mean_length": 201.9406295776367, "completions/mean_terminated_length": 201.9406295776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.030216981874157986, "epoch": 12.852157943067034, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2150859365.0, "reward": 8.192266368865967, "reward_std": 0.3060702055692673, "rewards/combined_reward/mean": 8.19226632118225, "rewards/combined_reward/std": 2.1787041783332826, "step": 21000 }, { "clip_ratio/high_max": 0.0018608584068715572, "clip_ratio/high_mean": 0.0011486702164499245, "clip_ratio/low_mean": 0.0007788341479075219, "clip_ratio/low_min": 0.00017812743778146493, "clip_ratio/region_mean": 0.0019275043397404565, "completions/clipped_ratio": 0.0, "completions/max_length": 615.9, "completions/max_terminated_length": 615.9, "completions/mean_length": 209.25234985351562, "completions/mean_terminated_length": 209.25234985351562, "completions/min_length": 53.8, "completions/min_terminated_length": 53.8, "entropy": 0.03019386241212487, "epoch": 12.870523415977962, "frac_reward_zero_std": 0.1583333395421505, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2153963054.0, "reward": 8.398930215835572, "reward_std": 0.2798346318304539, "rewards/combined_reward/mean": 8.398930311203003, "rewards/combined_reward/std": 1.7994829893112183, "step": 21030 }, { "clip_ratio/high_max": 0.0020694571920709375, "clip_ratio/high_mean": 0.0011915313831802148, "clip_ratio/low_mean": 0.0007161276809104796, "clip_ratio/low_min": 0.0002436246128733425, "clip_ratio/region_mean": 0.0019076590756109606, "completions/clipped_ratio": 0.0, "completions/max_length": 571.2, "completions/max_terminated_length": 571.2, "completions/mean_length": 210.49974517822267, "completions/mean_terminated_length": 210.49974517822267, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.029093335832779607, "epoch": 12.88888888888889, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.0262451171875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2157071533.0, "reward": 8.162282371520996, "reward_std": 0.368923220038414, "rewards/combined_reward/mean": 8.162282037734986, "rewards/combined_reward/std": 2.1968242168426513, "step": 21060 }, { "clip_ratio/high_max": 0.0018150433325596774, "clip_ratio/high_mean": 0.0010665141649951692, "clip_ratio/low_mean": 0.0007739772100952299, "clip_ratio/low_min": 0.00020533738376495118, "clip_ratio/region_mean": 0.001840491384791676, "completions/clipped_ratio": 0.0, "completions/max_length": 648.6, "completions/max_terminated_length": 648.6, "completions/mean_length": 207.64401245117188, "completions/mean_terminated_length": 207.64401245117188, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.027182147822653255, "epoch": 12.907254361799817, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.007659912109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2160169046.0, "reward": 8.081668329238891, "reward_std": 0.3339558571577072, "rewards/combined_reward/mean": 8.081668043136597, "rewards/combined_reward/std": 1.9528318285942077, "step": 21090 }, { "clip_ratio/high_max": 0.0016156620641898675, "clip_ratio/high_mean": 0.0009266409368137829, "clip_ratio/low_mean": 0.0008230487298836427, "clip_ratio/low_min": 0.00022426099021686242, "clip_ratio/region_mean": 0.0017496896699109735, "completions/clipped_ratio": 0.0, "completions/max_length": 600.4, "completions/max_terminated_length": 600.4, "completions/mean_length": 192.85937957763673, "completions/mean_terminated_length": 192.85937957763673, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.03134291944249223, "epoch": 12.925619834710744, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.00537109375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 2163209786.0, "reward": 8.25234055519104, "reward_std": 0.24539354741573333, "rewards/combined_reward/mean": 8.252340269088744, "rewards/combined_reward/std": 2.0251747608184814, "step": 21120 }, { "clip_ratio/high_max": 0.0021209733948732417, "clip_ratio/high_mean": 0.0011652849764989999, "clip_ratio/low_mean": 0.0008191130099779305, "clip_ratio/low_min": 0.0002785247915502017, "clip_ratio/region_mean": 0.001984397984536675, "completions/clipped_ratio": 0.0, "completions/max_length": 626.9, "completions/max_terminated_length": 626.9, "completions/mean_length": 202.57109985351562, "completions/mean_terminated_length": 202.57109985351562, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.028999118646606804, "epoch": 12.943985307621672, "frac_reward_zero_std": 0.23750000521540643, "grad_norm": 0.0301513671875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 2166287819.0, "reward": 7.830552482604981, "reward_std": 0.32754153311252593, "rewards/combined_reward/mean": 7.8305525302886965, "rewards/combined_reward/std": 2.6083951711654665, "step": 21150 }, { "clip_ratio/high_max": 0.0019698841705879508, "clip_ratio/high_mean": 0.0011218894932729503, "clip_ratio/low_mean": 0.0008396304417450059, "clip_ratio/low_min": 0.0001994445723539684, "clip_ratio/region_mean": 0.0019615199319863072, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 200.1893280029297, "completions/mean_terminated_length": 200.1893280029297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028692500619217753, "epoch": 12.9623507805326, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2169356706.0, "reward": 7.974457168579102, "reward_std": 0.30286143347620964, "rewards/combined_reward/mean": 7.974456691741944, "rewards/combined_reward/std": 2.372536301612854, "step": 21180 }, { "clip_ratio/high_max": 0.0014633102640800645, "clip_ratio/high_mean": 0.0008145068724843441, "clip_ratio/low_mean": 0.0007108086938387714, "clip_ratio/low_min": 0.00019471625128062442, "clip_ratio/region_mean": 0.0015253155540752535, "completions/clipped_ratio": 0.0, "completions/max_length": 610.8, "completions/max_terminated_length": 610.8, "completions/mean_length": 198.5914093017578, "completions/mean_terminated_length": 198.5914093017578, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.02519017681479454, "epoch": 12.980716253443527, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2172419457.0, "reward": 8.07041039466858, "reward_std": 0.2803119823336601, "rewards/combined_reward/mean": 8.070410108566284, "rewards/combined_reward/std": 2.145056480169296, "step": 21210 }, { "clip_ratio/high_max": 0.001629440802692746, "clip_ratio/high_mean": 0.0009174222479487071, "clip_ratio/low_mean": 0.0006996637808091084, "clip_ratio/low_min": 0.00021974065190685603, "clip_ratio/region_mean": 0.0016170860353061775, "completions/clipped_ratio": 0.0, "completions/max_length": 584.8, "completions/max_terminated_length": 584.8, "completions/mean_length": 201.8854187011719, "completions/mean_terminated_length": 201.8854187011719, "completions/min_length": 48.4, "completions/min_terminated_length": 48.4, "entropy": 0.025629117293283342, "epoch": 12.999081726354454, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2175494857.0, "reward": 8.17462158203125, "reward_std": 0.3122540920972824, "rewards/combined_reward/mean": 8.174621295928954, "rewards/combined_reward/std": 2.2563442826271056, "step": 21240 }, { "clip_ratio/high_max": 0.0024798841961873237, "clip_ratio/high_mean": 0.0014223125827647592, "clip_ratio/low_mean": 0.0011734958391228202, "clip_ratio/low_min": 0.0004436413318082904, "clip_ratio/region_mean": 0.002595808430237039, "completions/clipped_ratio": 0.0, "completions/max_length": 526.5454545454545, "completions/max_terminated_length": 526.5454545454545, "completions/mean_length": 198.0712654807351, "completions/mean_terminated_length": 198.0712654807351, "completions/min_length": 41.90909090909091, "completions/min_terminated_length": 41.90909090909091, "entropy": 0.029594042307895717, "epoch": 13.017753290480563, "frac_reward_zero_std": 0.185606065121564, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2178861686.0, "reward": 8.091102946888316, "reward_std": 0.29738457907329907, "rewards/combined_reward/mean": 8.091102600097656, "rewards/combined_reward/std": 2.2198882048780266, "step": 21270 }, { "clip_ratio/high_max": 0.002471395210159244, "clip_ratio/high_mean": 0.0015130740465489603, "clip_ratio/low_mean": 0.0009801648997381563, "clip_ratio/low_min": 0.00028168151790547804, "clip_ratio/region_mean": 0.002493238924944308, "completions/clipped_ratio": 0.0, "completions/max_length": 544.4, "completions/max_terminated_length": 544.4, "completions/mean_length": 213.59870452880858, "completions/mean_terminated_length": 213.59870452880858, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.032942129958731435, "epoch": 13.03611876339149, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2181982065.0, "reward": 8.052363443374634, "reward_std": 0.3426604583859444, "rewards/combined_reward/mean": 8.052363157272339, "rewards/combined_reward/std": 2.2902815699577332, "step": 21300 }, { "clip_ratio/high_max": 0.002624961700833713, "clip_ratio/high_mean": 0.0014518934998098605, "clip_ratio/low_mean": 0.0009189090502331965, "clip_ratio/low_min": 0.0002716815239788654, "clip_ratio/region_mean": 0.0023708025614420573, "completions/clipped_ratio": 0.0, "completions/max_length": 532.4, "completions/max_terminated_length": 532.4, "completions/mean_length": 204.4213607788086, "completions/mean_terminated_length": 204.4213607788086, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.027682192934056122, "epoch": 13.054484236302418, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2185067203.0, "reward": 8.214543581008911, "reward_std": 0.2699645161628723, "rewards/combined_reward/mean": 8.214543437957763, "rewards/combined_reward/std": 1.9859506249427796, "step": 21330 }, { "clip_ratio/high_max": 0.002259181115126315, "clip_ratio/high_mean": 0.0013722773765039165, "clip_ratio/low_mean": 0.000978956632449505, "clip_ratio/low_min": 0.00030738427109705904, "clip_ratio/region_mean": 0.0023512340145922885, "completions/clipped_ratio": 0.0, "completions/max_length": 547.2, "completions/max_terminated_length": 547.2, "completions/mean_length": 198.72370300292968, "completions/mean_terminated_length": 198.72370300292968, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02841983417359491, "epoch": 13.072849709213346, "frac_reward_zero_std": 0.1333333369344473, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2188130462.0, "reward": 8.244747161865234, "reward_std": 0.3287704274058342, "rewards/combined_reward/mean": 8.244746923446655, "rewards/combined_reward/std": 2.187513267993927, "step": 21360 }, { "clip_ratio/high_max": 0.002560660044158188, "clip_ratio/high_mean": 0.0014315741677516296, "clip_ratio/low_mean": 0.0008806960006040754, "clip_ratio/low_min": 0.00017582060560622875, "clip_ratio/region_mean": 0.002312270176965588, "completions/clipped_ratio": 0.0, "completions/max_length": 664.8, "completions/max_terminated_length": 664.8, "completions/mean_length": 207.06693115234376, "completions/mean_terminated_length": 207.06693115234376, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02834665374830365, "epoch": 13.091215182124273, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2191225759.0, "reward": 8.153453874588013, "reward_std": 0.24907680004835128, "rewards/combined_reward/mean": 8.153453683853149, "rewards/combined_reward/std": 1.9728279829025268, "step": 21390 }, { "clip_ratio/high_max": 0.0023616562563499124, "clip_ratio/high_mean": 0.001383015994724701, "clip_ratio/low_mean": 0.0010392487598437582, "clip_ratio/low_min": 0.00031750159323564733, "clip_ratio/region_mean": 0.0024222647722732895, "completions/clipped_ratio": 0.0, "completions/max_length": 634.7, "completions/max_terminated_length": 634.7, "completions/mean_length": 204.7166717529297, "completions/mean_terminated_length": 204.7166717529297, "completions/min_length": 47.2, "completions/min_terminated_length": 47.2, "entropy": 0.030596214812248945, "epoch": 13.1095806550352, "frac_reward_zero_std": 0.19583334103226663, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2194312031.0, "reward": 7.792387247085571, "reward_std": 0.32801480293273927, "rewards/combined_reward/mean": 7.7923869609832765, "rewards/combined_reward/std": 2.4662142157554627, "step": 21420 }, { "clip_ratio/high_max": 0.0026088342929142527, "clip_ratio/high_mean": 0.0014224844291068925, "clip_ratio/low_mean": 0.0009638740578642077, "clip_ratio/low_min": 0.0002439825196537034, "clip_ratio/region_mean": 0.0023863584884869244, "completions/clipped_ratio": 0.0, "completions/max_length": 619.4, "completions/max_terminated_length": 619.4, "completions/mean_length": 205.52396240234376, "completions/mean_terminated_length": 205.52396240234376, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.031063207983970643, "epoch": 13.127946127946128, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2197401403.0, "reward": 8.115276622772218, "reward_std": 0.34436977207660674, "rewards/combined_reward/mean": 8.115276384353638, "rewards/combined_reward/std": 2.33791036605835, "step": 21450 }, { "clip_ratio/high_max": 0.002017099372460507, "clip_ratio/high_mean": 0.001129393364074834, "clip_ratio/low_mean": 0.0010087324495543725, "clip_ratio/low_min": 0.0002684072440994593, "clip_ratio/region_mean": 0.0021381258268471964, "completions/clipped_ratio": 0.0, "completions/max_length": 594.2, "completions/max_terminated_length": 594.2, "completions/mean_length": 193.23802642822267, "completions/mean_terminated_length": 193.23802642822267, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028053884161636233, "epoch": 13.146311600857056, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2200443597.0, "reward": 7.948208045959473, "reward_std": 0.3041040077805519, "rewards/combined_reward/mean": 7.948208141326904, "rewards/combined_reward/std": 2.3363004326820374, "step": 21480 }, { "clip_ratio/high_max": 0.0023514403588099715, "clip_ratio/high_mean": 0.0013189981514490985, "clip_ratio/low_mean": 0.0008375576510540365, "clip_ratio/low_min": 0.00023474493694569293, "clip_ratio/region_mean": 0.0021565558063836457, "completions/clipped_ratio": 0.0, "completions/max_length": 584.3, "completions/max_terminated_length": 584.3, "completions/mean_length": 206.30938110351562, "completions/mean_terminated_length": 206.30938110351562, "completions/min_length": 46.7, "completions/min_terminated_length": 46.7, "entropy": 0.030604278125489753, "epoch": 13.164677073767983, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2203535985.0, "reward": 8.10278754234314, "reward_std": 0.33867553621530533, "rewards/combined_reward/mean": 8.102787399291993, "rewards/combined_reward/std": 2.1747920632362367, "step": 21510 }, { "clip_ratio/high_max": 0.00196357160093612, "clip_ratio/high_mean": 0.0011238255620507212, "clip_ratio/low_mean": 0.0008482618159784276, "clip_ratio/low_min": 0.000206892875576159, "clip_ratio/region_mean": 0.001972087377483452, "completions/clipped_ratio": 0.0, "completions/max_length": 545.1, "completions/max_terminated_length": 545.1, "completions/mean_length": 200.03333892822266, "completions/mean_terminated_length": 200.03333892822266, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02493072294164449, "epoch": 13.18304254667891, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.0150146484375, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 2206604273.0, "reward": 8.156020593643188, "reward_std": 0.2394617848098278, "rewards/combined_reward/mean": 8.156020498275756, "rewards/combined_reward/std": 2.212590742111206, "step": 21540 }, { "clip_ratio/high_max": 0.0027433936619975915, "clip_ratio/high_mean": 0.0015767860364576335, "clip_ratio/low_mean": 0.0009332293101275961, "clip_ratio/low_min": 0.0002175402665064515, "clip_ratio/region_mean": 0.0025100153463426976, "completions/clipped_ratio": 0.0, "completions/max_length": 596.6, "completions/max_terminated_length": 596.6, "completions/mean_length": 193.99557647705078, "completions/mean_terminated_length": 193.99557647705078, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.03133858895550171, "epoch": 13.201408019589838, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2209649376.0, "reward": 7.960494422912598, "reward_std": 0.2640508085489273, "rewards/combined_reward/mean": 7.9604939937591555, "rewards/combined_reward/std": 2.3391681790351866, "step": 21570 }, { "clip_ratio/high_max": 0.0019994839080027305, "clip_ratio/high_mean": 0.0011465983994033498, "clip_ratio/low_mean": 0.0009682283215321756, "clip_ratio/low_min": 0.00028500275414747496, "clip_ratio/region_mean": 0.0021148267119618443, "completions/clipped_ratio": 0.0, "completions/max_length": 609.8, "completions/max_terminated_length": 609.8, "completions/mean_length": 199.76067962646485, "completions/mean_terminated_length": 199.76067962646485, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027315803446496527, "epoch": 13.219773492500765, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2212716617.0, "reward": 8.00980634689331, "reward_std": 0.2785539448261261, "rewards/combined_reward/mean": 8.00980610847473, "rewards/combined_reward/std": 2.4695659637451173, "step": 21600 }, { "clip_ratio/high_max": 0.0027934483650218075, "clip_ratio/high_mean": 0.001547048032322588, "clip_ratio/low_mean": 0.0010486611721717055, "clip_ratio/low_min": 0.00024759188017924315, "clip_ratio/region_mean": 0.0025957092104363253, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 197.41068115234376, "completions/mean_terminated_length": 197.41068115234376, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.03324878327548504, "epoch": 13.238138965411693, "frac_reward_zero_std": 0.1375000037252903, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2215774834.0, "reward": 8.201745700836181, "reward_std": 0.28946367278695107, "rewards/combined_reward/mean": 8.201745510101318, "rewards/combined_reward/std": 1.9923046350479126, "step": 21630 }, { "clip_ratio/high_max": 0.0023953270448449377, "clip_ratio/high_mean": 0.0013884345081654223, "clip_ratio/low_mean": 0.0009072862862619028, "clip_ratio/low_min": 0.0002964294554355244, "clip_ratio/region_mean": 0.0022957207787840162, "completions/clipped_ratio": 0.0, "completions/max_length": 579.7, "completions/max_terminated_length": 579.7, "completions/mean_length": 206.2203170776367, "completions/mean_terminated_length": 206.2203170776367, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.0268695961839209, "epoch": 13.25650443832262, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2218866880.0, "reward": 8.13909978866577, "reward_std": 0.29240331053733826, "rewards/combined_reward/mean": 8.139099359512329, "rewards/combined_reward/std": 2.1874403834342955, "step": 21660 }, { "clip_ratio/high_max": 0.0023924121977567363, "clip_ratio/high_mean": 0.0013528267501290733, "clip_ratio/low_mean": 0.0008708810846049649, "clip_ratio/low_min": 0.00021202661616068024, "clip_ratio/region_mean": 0.002223707810723378, "completions/clipped_ratio": 0.0, "completions/max_length": 535.6, "completions/max_terminated_length": 535.6, "completions/mean_length": 189.93958740234376, "completions/mean_terminated_length": 189.93958740234376, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.03079785071313381, "epoch": 13.274869911233548, "frac_reward_zero_std": 0.20833334214985372, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2221896408.0, "reward": 8.06919493675232, "reward_std": 0.26353218108415605, "rewards/combined_reward/mean": 8.06919460296631, "rewards/combined_reward/std": 2.192227578163147, "step": 21690 }, { "clip_ratio/high_max": 0.0019261000479067055, "clip_ratio/high_mean": 0.001027912630646218, "clip_ratio/low_mean": 0.0008692277640875545, "clip_ratio/low_min": 0.0001963879134564195, "clip_ratio/region_mean": 0.0018971403910351607, "completions/clipped_ratio": 0.0, "completions/max_length": 687.2, "completions/max_terminated_length": 687.2, "completions/mean_length": 198.3987030029297, "completions/mean_terminated_length": 198.3987030029297, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.025323556433431805, "epoch": 13.293235384144475, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2224958419.0, "reward": 7.934550142288208, "reward_std": 0.30382662266492844, "rewards/combined_reward/mean": 7.934549760818482, "rewards/combined_reward/std": 2.3591052532196044, "step": 21720 }, { "clip_ratio/high_max": 0.002334110021668797, "clip_ratio/high_mean": 0.0012605207518693836, "clip_ratio/low_mean": 0.001017888481328555, "clip_ratio/low_min": 0.0002940630285593215, "clip_ratio/region_mean": 0.0022784092390793374, "completions/clipped_ratio": 0.0, "completions/max_length": 599.4, "completions/max_terminated_length": 599.4, "completions/mean_length": 199.7213607788086, "completions/mean_terminated_length": 199.7213607788086, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.028237455058842897, "epoch": 13.311600857055403, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2228025509.0, "reward": 8.219541597366334, "reward_std": 0.3636612206697464, "rewards/combined_reward/mean": 8.219541311264038, "rewards/combined_reward/std": 2.078858804702759, "step": 21750 }, { "clip_ratio/high_max": 0.0025141973649927727, "clip_ratio/high_mean": 0.0015084013831559182, "clip_ratio/low_mean": 0.0009995944114052691, "clip_ratio/low_min": 0.0003020751729006103, "clip_ratio/region_mean": 0.0025079957650935587, "completions/clipped_ratio": 0.0, "completions/max_length": 585.7, "completions/max_terminated_length": 585.7, "completions/mean_length": 200.4875061035156, "completions/mean_terminated_length": 200.4875061035156, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.032872663081313175, "epoch": 13.32996632996633, "frac_reward_zero_std": 0.13333333805203437, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2231095541.0, "reward": 8.239881134033203, "reward_std": 0.34078790694475175, "rewards/combined_reward/mean": 8.23988094329834, "rewards/combined_reward/std": 2.1891435503959658, "step": 21780 }, { "clip_ratio/high_max": 0.002441550808240815, "clip_ratio/high_mean": 0.001443979907950658, "clip_ratio/low_mean": 0.0010102373295618843, "clip_ratio/low_min": 0.00028267845312560287, "clip_ratio/region_mean": 0.002454217229387723, "completions/clipped_ratio": 0.0, "completions/max_length": 670.9, "completions/max_terminated_length": 670.9, "completions/mean_length": 212.08490295410155, "completions/mean_terminated_length": 212.08490295410155, "completions/min_length": 48.6, "completions/min_terminated_length": 48.6, "entropy": 0.03051161045829455, "epoch": 13.348331802877258, "frac_reward_zero_std": 0.17916667200624942, "grad_norm": 0.005126953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2234210107.0, "reward": 7.945458889007568, "reward_std": 0.28841356933116913, "rewards/combined_reward/mean": 7.945458793640137, "rewards/combined_reward/std": 2.435744059085846, "step": 21810 }, { "clip_ratio/high_max": 0.002587857328762766, "clip_ratio/high_mean": 0.0014191507936629933, "clip_ratio/low_mean": 0.0008947830728478341, "clip_ratio/low_min": 0.00023258683116485676, "clip_ratio/region_mean": 0.0023139338668746253, "completions/clipped_ratio": 0.0, "completions/max_length": 554.5, "completions/max_terminated_length": 554.5, "completions/mean_length": 199.57865142822266, "completions/mean_terminated_length": 199.57865142822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027936623555918534, "epoch": 13.366697275788185, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0037994384765625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2237276649.0, "reward": 8.179902744293212, "reward_std": 0.280223585665226, "rewards/combined_reward/mean": 8.179902029037475, "rewards/combined_reward/std": 2.154122030735016, "step": 21840 }, { "clip_ratio/high_max": 0.002723189791625676, "clip_ratio/high_mean": 0.0015524167198843013, "clip_ratio/low_mean": 0.0011289218905100522, "clip_ratio/low_min": 0.00036766462532492977, "clip_ratio/region_mean": 0.002681338620459428, "completions/clipped_ratio": 0.0, "completions/max_length": 559.3, "completions/max_terminated_length": 559.3, "completions/mean_length": 198.8901107788086, "completions/mean_terminated_length": 198.8901107788086, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03066217016118268, "epoch": 13.385062748699113, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.0067138671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2240340547.0, "reward": 7.88869276046753, "reward_std": 0.328561332821846, "rewards/combined_reward/mean": 7.888692569732666, "rewards/combined_reward/std": 2.422518587112427, "step": 21870 }, { "clip_ratio/high_max": 0.002533778022916522, "clip_ratio/high_mean": 0.001418945176070944, "clip_ratio/low_mean": 0.0010719188556322478, "clip_ratio/low_min": 0.000383464931655908, "clip_ratio/region_mean": 0.0024908640300661017, "completions/clipped_ratio": 0.0, "completions/max_length": 564.2, "completions/max_terminated_length": 564.2, "completions/mean_length": 209.5362060546875, "completions/mean_terminated_length": 209.5362060546875, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.029465698908704024, "epoch": 13.40342822161004, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2243445326.0, "reward": 7.896484231948852, "reward_std": 0.3627794817090034, "rewards/combined_reward/mean": 7.896484136581421, "rewards/combined_reward/std": 2.3351905822753904, "step": 21900 }, { "clip_ratio/high_max": 0.002645598717208486, "clip_ratio/high_mean": 0.0014366788459786524, "clip_ratio/low_mean": 0.0009971638001540365, "clip_ratio/low_min": 0.000279752704227576, "clip_ratio/region_mean": 0.0024338426502557315, "completions/clipped_ratio": 0.0, "completions/max_length": 589.5, "completions/max_terminated_length": 589.5, "completions/mean_length": 195.39531707763672, "completions/mean_terminated_length": 195.39531707763672, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.029744734009727835, "epoch": 13.421793694520968, "frac_reward_zero_std": 0.17083334028720856, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2246495804.0, "reward": 7.859604072570801, "reward_std": 0.28155178278684617, "rewards/combined_reward/mean": 7.859603929519653, "rewards/combined_reward/std": 2.352794373035431, "step": 21930 }, { "clip_ratio/high_max": 0.0021238428763657186, "clip_ratio/high_mean": 0.0012527486758093195, "clip_ratio/low_mean": 0.0008122670182880635, "clip_ratio/low_min": 0.0002449883703472248, "clip_ratio/region_mean": 0.002065015700645745, "completions/clipped_ratio": 0.0, "completions/max_length": 612.8, "completions/max_terminated_length": 612.8, "completions/mean_length": 207.50625305175782, "completions/mean_terminated_length": 207.50625305175782, "completions/min_length": 47.4, "completions/min_terminated_length": 47.4, "entropy": 0.02579916696995497, "epoch": 13.440159167431895, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2249592788.0, "reward": 8.072193002700805, "reward_std": 0.29202996045351026, "rewards/combined_reward/mean": 8.072192859649657, "rewards/combined_reward/std": 2.52458860874176, "step": 21960 }, { "clip_ratio/high_max": 0.002838783418701496, "clip_ratio/high_mean": 0.0016347632605175021, "clip_ratio/low_mean": 0.000962747480540808, "clip_ratio/low_min": 0.0002884884306695312, "clip_ratio/region_mean": 0.0025975107317208314, "completions/clipped_ratio": 0.0, "completions/max_length": 658.9, "completions/max_terminated_length": 658.9, "completions/mean_length": 199.8010452270508, "completions/mean_terminated_length": 199.8010452270508, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03275629715062678, "epoch": 13.458524640342823, "frac_reward_zero_std": 0.2291666716337204, "grad_norm": 0.0250244140625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2252660184.0, "reward": 8.109378051757812, "reward_std": 0.3642192676663399, "rewards/combined_reward/mean": 8.109377479553222, "rewards/combined_reward/std": 2.1789112210273744, "step": 21990 }, { "clip_ratio/high_max": 0.002427420794265345, "clip_ratio/high_mean": 0.0014316690957154302, "clip_ratio/low_mean": 0.0010605965831928187, "clip_ratio/low_min": 0.00030314754003484267, "clip_ratio/region_mean": 0.0024922657001297923, "completions/clipped_ratio": 0.0, "completions/max_length": 502.2, "completions/max_terminated_length": 502.2, "completions/mean_length": 197.38021240234374, "completions/mean_terminated_length": 197.38021240234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02986131766034911, "epoch": 13.47689011325375, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2255718284.0, "reward": 7.976070785522461, "reward_std": 0.3662046268582344, "rewards/combined_reward/mean": 7.97607045173645, "rewards/combined_reward/std": 2.3339563846588134, "step": 22020 }, { "clip_ratio/high_max": 0.0023215602263614225, "clip_ratio/high_mean": 0.0013462523150034637, "clip_ratio/low_mean": 0.0009637212600258257, "clip_ratio/low_min": 0.00025496981494749584, "clip_ratio/region_mean": 0.0023099735832753745, "completions/clipped_ratio": 0.0, "completions/max_length": 698.1, "completions/max_terminated_length": 698.1, "completions/mean_length": 206.9250061035156, "completions/mean_terminated_length": 206.9250061035156, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028580171785627804, "epoch": 13.495255586164678, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2258813036.0, "reward": 8.02281928062439, "reward_std": 0.31235743761062623, "rewards/combined_reward/mean": 8.02281928062439, "rewards/combined_reward/std": 2.205497515201569, "step": 22050 }, { "clip_ratio/high_max": 0.00255560452399853, "clip_ratio/high_mean": 0.0015233652765649216, "clip_ratio/low_mean": 0.0009832949997265435, "clip_ratio/low_min": 0.0002568910720583517, "clip_ratio/region_mean": 0.00250666028005071, "completions/clipped_ratio": 0.0, "completions/max_length": 614.7, "completions/max_terminated_length": 614.7, "completions/mean_length": 206.4562545776367, "completions/mean_terminated_length": 206.4562545776367, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.031152904514844218, "epoch": 13.513621059075604, "frac_reward_zero_std": 0.12500000409781933, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2261905988.0, "reward": 7.961917781829834, "reward_std": 0.3967642650008202, "rewards/combined_reward/mean": 7.961917495727539, "rewards/combined_reward/std": 2.233812081813812, "step": 22080 }, { "clip_ratio/high_max": 0.002462044035200961, "clip_ratio/high_mean": 0.001390911891212454, "clip_ratio/low_mean": 0.0009139052966929739, "clip_ratio/low_min": 0.0003160416415388075, "clip_ratio/region_mean": 0.0023048171953026515, "completions/clipped_ratio": 0.0, "completions/max_length": 509.7, "completions/max_terminated_length": 509.7, "completions/mean_length": 200.86953735351562, "completions/mean_terminated_length": 200.86953735351562, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.027775520109571517, "epoch": 13.531986531986533, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2264977487.0, "reward": 7.79345326423645, "reward_std": 0.38266296982765197, "rewards/combined_reward/mean": 7.793452978134155, "rewards/combined_reward/std": 2.5954743385314942, "step": 22110 }, { "clip_ratio/high_max": 0.002040960971013798, "clip_ratio/high_mean": 0.0011695707419373017, "clip_ratio/low_mean": 0.0009208782635444853, "clip_ratio/low_min": 0.00023289009186555631, "clip_ratio/region_mean": 0.0020904490035415315, "completions/clipped_ratio": 0.0, "completions/max_length": 515.7, "completions/max_terminated_length": 515.7, "completions/mean_length": 198.7619842529297, "completions/mean_terminated_length": 198.7619842529297, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.025459798130517203, "epoch": 13.550352004897459, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2268040893.0, "reward": 7.7956929206848145, "reward_std": 0.3871389254927635, "rewards/combined_reward/mean": 7.795692825317383, "rewards/combined_reward/std": 2.5441155791282655, "step": 22140 }, { "clip_ratio/high_max": 0.0025877301193152863, "clip_ratio/high_mean": 0.0016173546835489105, "clip_ratio/low_mean": 0.0009257785614560514, "clip_ratio/low_min": 0.00022758231789339335, "clip_ratio/region_mean": 0.002543133232393302, "completions/clipped_ratio": 0.0, "completions/max_length": 581.1, "completions/max_terminated_length": 581.1, "completions/mean_length": 208.67578735351563, "completions/mean_terminated_length": 208.67578735351563, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.03197816121391952, "epoch": 13.568717477808386, "frac_reward_zero_std": 0.1375000037252903, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2271142368.0, "reward": 8.046406126022339, "reward_std": 0.3839550852775574, "rewards/combined_reward/mean": 8.04640564918518, "rewards/combined_reward/std": 2.191130447387695, "step": 22170 }, { "clip_ratio/high_max": 0.0026123379973190214, "clip_ratio/high_mean": 0.0014395802475822468, "clip_ratio/low_mean": 0.0008723091893140614, "clip_ratio/low_min": 0.00023973506104084662, "clip_ratio/region_mean": 0.002311889434349723, "completions/clipped_ratio": 0.0, "completions/max_length": 512.5, "completions/max_terminated_length": 512.5, "completions/mean_length": 188.31693115234376, "completions/mean_terminated_length": 188.31693115234376, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028728798124939205, "epoch": 13.587082950719314, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2274165665.0, "reward": 8.068527603149414, "reward_std": 0.30636862963438033, "rewards/combined_reward/mean": 8.068527603149414, "rewards/combined_reward/std": 2.2738974571228026, "step": 22200 }, { "clip_ratio/high_max": 0.002150965332596873, "clip_ratio/high_mean": 0.0012886982407508186, "clip_ratio/low_mean": 0.0009716967358447922, "clip_ratio/low_min": 0.000264775180403376, "clip_ratio/region_mean": 0.0022603949861756214, "completions/clipped_ratio": 0.0, "completions/max_length": 597.9, "completions/max_terminated_length": 597.9, "completions/mean_length": 207.47448272705077, "completions/mean_terminated_length": 207.47448272705077, "completions/min_length": 50.3, "completions/min_terminated_length": 50.3, "entropy": 0.029299131866234045, "epoch": 13.605448423630241, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2277262527.0, "reward": 8.16000919342041, "reward_std": 0.4172964975237846, "rewards/combined_reward/mean": 8.160008955001832, "rewards/combined_reward/std": 2.2145231485366823, "step": 22230 }, { "clip_ratio/high_max": 0.0025558472184153895, "clip_ratio/high_mean": 0.0014743932571946062, "clip_ratio/low_mean": 0.0009936849367174242, "clip_ratio/low_min": 0.0002150886386516504, "clip_ratio/region_mean": 0.0024680781959129186, "completions/clipped_ratio": 0.0, "completions/max_length": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 194.1697967529297, "completions/mean_terminated_length": 194.1697967529297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029555474369165797, "epoch": 13.623813896541169, "frac_reward_zero_std": 0.2208333432674408, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2280308299.0, "reward": 7.835624742507934, "reward_std": 0.2980874627828598, "rewards/combined_reward/mean": 7.835624694824219, "rewards/combined_reward/std": 2.3773970723152162, "step": 22260 }, { "clip_ratio/high_max": 0.002878593688850136, "clip_ratio/high_mean": 0.0015747088236821583, "clip_ratio/low_mean": 0.001121803374432299, "clip_ratio/low_min": 0.00037088195046332354, "clip_ratio/region_mean": 0.0026965121828349463, "completions/clipped_ratio": 0.0, "completions/max_length": 596.2, "completions/max_terminated_length": 596.2, "completions/mean_length": 190.32474365234376, "completions/mean_terminated_length": 190.32474365234376, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.03091568952271094, "epoch": 13.642179369452096, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.0057373046875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2283339306.0, "reward": 7.94985613822937, "reward_std": 0.33214312344789504, "rewards/combined_reward/mean": 7.949855947494507, "rewards/combined_reward/std": 2.3447606801986693, "step": 22290 }, { "clip_ratio/high_max": 0.0023680717931711115, "clip_ratio/high_mean": 0.0013739947828677638, "clip_ratio/low_mean": 0.0009025176671760468, "clip_ratio/low_min": 0.0002538060053969578, "clip_ratio/region_mean": 0.0022765124347642997, "completions/clipped_ratio": 0.0, "completions/max_length": 581.4, "completions/max_terminated_length": 581.4, "completions/mean_length": 186.73334045410155, "completions/mean_terminated_length": 186.73334045410155, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.02674944542037944, "epoch": 13.660544842363024, "frac_reward_zero_std": 0.18750000298023223, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2286356522.0, "reward": 7.819571781158447, "reward_std": 0.37088376730680467, "rewards/combined_reward/mean": 7.8195716381073, "rewards/combined_reward/std": 2.3771053314208985, "step": 22320 }, { "clip_ratio/high_max": 0.002198319631861523, "clip_ratio/high_mean": 0.0012994694251877567, "clip_ratio/low_mean": 0.0009795881140841325, "clip_ratio/low_min": 0.0003184862131699144, "clip_ratio/region_mean": 0.002279057558916975, "completions/clipped_ratio": 0.0, "completions/max_length": 525.2, "completions/max_terminated_length": 525.2, "completions/mean_length": 201.2151077270508, "completions/mean_terminated_length": 201.2151077270508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027399774407967925, "epoch": 13.678910315273951, "frac_reward_zero_std": 0.13750000149011612, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2289429348.0, "reward": 8.034841775894165, "reward_std": 0.3592576593160629, "rewards/combined_reward/mean": 8.034841394424438, "rewards/combined_reward/std": 2.2076780915260317, "step": 22350 }, { "clip_ratio/high_max": 0.00267687795664339, "clip_ratio/high_mean": 0.0014550550815329189, "clip_ratio/low_mean": 0.0009440307237431019, "clip_ratio/low_min": 0.00017022991669364275, "clip_ratio/region_mean": 0.002399085786601063, "completions/clipped_ratio": 0.0, "completions/max_length": 504.9, "completions/max_terminated_length": 504.9, "completions/mean_length": 187.9796936035156, "completions/mean_terminated_length": 187.9796936035156, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.027929604037975272, "epoch": 13.697275788184879, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.026611328125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2292451350.0, "reward": 8.088045930862426, "reward_std": 0.34712214320898055, "rewards/combined_reward/mean": 8.088045644760133, "rewards/combined_reward/std": 2.2541550397872925, "step": 22380 }, { "clip_ratio/high_max": 0.0024360937968594953, "clip_ratio/high_mean": 0.0013784975509527916, "clip_ratio/low_mean": 0.0009889269473205785, "clip_ratio/low_min": 0.00026788543691509404, "clip_ratio/region_mean": 0.002367424514765541, "completions/clipped_ratio": 0.0, "completions/max_length": 639.8, "completions/max_terminated_length": 639.8, "completions/mean_length": 193.2369842529297, "completions/mean_terminated_length": 193.2369842529297, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02822124205219249, "epoch": 13.715641261095806, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2295493540.0, "reward": 7.87997784614563, "reward_std": 0.34086680263280866, "rewards/combined_reward/mean": 7.879977655410767, "rewards/combined_reward/std": 2.433146071434021, "step": 22410 }, { "clip_ratio/high_max": 0.0027403252584917936, "clip_ratio/high_mean": 0.001443540288407045, "clip_ratio/low_mean": 0.0011924863943325666, "clip_ratio/low_min": 0.00047178125920860716, "clip_ratio/region_mean": 0.0026360266681876967, "completions/clipped_ratio": 0.0, "completions/max_length": 558.2, "completions/max_terminated_length": 558.2, "completions/mean_length": 200.43047332763672, "completions/mean_terminated_length": 200.43047332763672, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.030373442002261677, "epoch": 13.734006734006734, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2298563353.0, "reward": 8.06602077484131, "reward_std": 0.3421484440565109, "rewards/combined_reward/mean": 8.066020631790161, "rewards/combined_reward/std": 2.2934502601623534, "step": 22440 }, { "clip_ratio/high_max": 0.002269382568677732, "clip_ratio/high_mean": 0.001383226611505961, "clip_ratio/low_mean": 0.0009755624955384216, "clip_ratio/low_min": 0.00027562986506382006, "clip_ratio/region_mean": 0.0023587891021937444, "completions/clipped_ratio": 0.0, "completions/max_length": 571.4, "completions/max_terminated_length": 571.4, "completions/mean_length": 193.61641387939454, "completions/mean_terminated_length": 193.61641387939454, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.029402020666748286, "epoch": 13.752372206917661, "frac_reward_zero_std": 0.20000000894069672, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2301607000.0, "reward": 7.784182596206665, "reward_std": 0.376604799926281, "rewards/combined_reward/mean": 7.784182214736939, "rewards/combined_reward/std": 2.5764938473701475, "step": 22470 }, { "clip_ratio/high_max": 0.0023854887801765773, "clip_ratio/high_mean": 0.001263026656791529, "clip_ratio/low_mean": 0.0008813108626175866, "clip_ratio/low_min": 0.00019046812133941178, "clip_ratio/region_mean": 0.002144337512921387, "completions/clipped_ratio": 0.0, "completions/max_length": 559.1, "completions/max_terminated_length": 559.1, "completions/mean_length": 197.59115142822264, "completions/mean_terminated_length": 197.59115142822264, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028161256558572254, "epoch": 13.770737679828589, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2304665910.0, "reward": 8.15977201461792, "reward_std": 0.305799075961113, "rewards/combined_reward/mean": 8.159771633148193, "rewards/combined_reward/std": 2.165465772151947, "step": 22500 }, { "clip_ratio/high_max": 0.0026631259635905733, "clip_ratio/high_mean": 0.0015443295322863074, "clip_ratio/low_mean": 0.0009034413180415868, "clip_ratio/low_min": 0.0001997780588377888, "clip_ratio/region_mean": 0.002447770865789304, "completions/clipped_ratio": 0.0, "completions/max_length": 578.5, "completions/max_terminated_length": 578.5, "completions/mean_length": 198.45912017822266, "completions/mean_terminated_length": 198.45912017822266, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028707667191823324, "epoch": 13.789103152739516, "frac_reward_zero_std": 0.2125000074505806, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2307728153.0, "reward": 8.017341375350952, "reward_std": 0.33176025152206423, "rewards/combined_reward/mean": 8.017341089248657, "rewards/combined_reward/std": 2.355319023132324, "step": 22530 }, { "clip_ratio/high_max": 0.0024200153529818635, "clip_ratio/high_mean": 0.0013555655799185236, "clip_ratio/low_mean": 0.0010424162172663878, "clip_ratio/low_min": 0.00033889829986340675, "clip_ratio/region_mean": 0.0023979817973061777, "completions/clipped_ratio": 0.0, "completions/max_length": 594.8, "completions/max_terminated_length": 594.8, "completions/mean_length": 195.93307800292968, "completions/mean_terminated_length": 195.93307800292968, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.029875248080740373, "epoch": 13.807468625650444, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.0103759765625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2310780696.0, "reward": 8.106049203872681, "reward_std": 0.33631465435028074, "rewards/combined_reward/mean": 8.106048822402954, "rewards/combined_reward/std": 2.1959927558898924, "step": 22560 }, { "clip_ratio/high_max": 0.0024815598197164947, "clip_ratio/high_mean": 0.0014759013851289638, "clip_ratio/low_mean": 0.0009967573764394426, "clip_ratio/low_min": 0.0002645576908738197, "clip_ratio/region_mean": 0.002472658734768629, "completions/clipped_ratio": 0.0, "completions/max_length": 573.6, "completions/max_terminated_length": 573.6, "completions/mean_length": 211.6408905029297, "completions/mean_terminated_length": 211.6408905029297, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.030787319852970542, "epoch": 13.825834098561371, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2313893557.0, "reward": 7.909446287155151, "reward_std": 0.42189075946807864, "rewards/combined_reward/mean": 7.9094460010528564, "rewards/combined_reward/std": 2.2635799646377563, "step": 22590 }, { "clip_ratio/high_max": 0.002712577787072708, "clip_ratio/high_mean": 0.0014937136967394812, "clip_ratio/low_mean": 0.0009525863545907972, "clip_ratio/low_min": 0.00022429032178479248, "clip_ratio/region_mean": 0.0024463000687925764, "completions/clipped_ratio": 0.0, "completions/max_length": 582.5, "completions/max_terminated_length": 582.5, "completions/mean_length": 203.9367233276367, "completions/mean_terminated_length": 203.9367233276367, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.030814486602321268, "epoch": 13.844199571472299, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.0123291015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2316976834.0, "reward": 8.129057264328003, "reward_std": 0.38800006061792375, "rewards/combined_reward/mean": 8.129057025909423, "rewards/combined_reward/std": 2.5015074133872988, "step": 22620 }, { "clip_ratio/high_max": 0.0025088601129633994, "clip_ratio/high_mean": 0.0014159074972364275, "clip_ratio/low_mean": 0.000946029825960674, "clip_ratio/low_min": 0.000257999739793983, "clip_ratio/region_mean": 0.002361937312525697, "completions/clipped_ratio": 0.0, "completions/max_length": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 194.3424514770508, "completions/mean_terminated_length": 194.3424514770508, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.029676352394744754, "epoch": 13.862565044383226, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2320023269.0, "reward": 7.953255319595337, "reward_std": 0.36672686487436296, "rewards/combined_reward/mean": 7.953255033493042, "rewards/combined_reward/std": 2.5313905119895934, "step": 22650 }, { "clip_ratio/high_max": 0.0029661470723416033, "clip_ratio/high_mean": 0.001717961007913497, "clip_ratio/low_mean": 0.0010884596763692874, "clip_ratio/low_min": 0.00033990913192004276, "clip_ratio/region_mean": 0.002806420694105327, "completions/clipped_ratio": 0.0, "completions/max_length": 637.7, "completions/max_terminated_length": 637.7, "completions/mean_length": 191.4830764770508, "completions/mean_terminated_length": 191.4830764770508, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03133100040722638, "epoch": 13.880930517294154, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2323058724.0, "reward": 8.017948627471924, "reward_std": 0.34140867218375204, "rewards/combined_reward/mean": 8.017948532104493, "rewards/combined_reward/std": 2.3308137893676757, "step": 22680 }, { "clip_ratio/high_max": 0.0025105236512899865, "clip_ratio/high_mean": 0.001417396169138859, "clip_ratio/low_mean": 0.0009334699089473968, "clip_ratio/low_min": 0.0002551469170915273, "clip_ratio/region_mean": 0.0023508660810572716, "completions/clipped_ratio": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 192.18724670410157, "completions/mean_terminated_length": 192.18724670410157, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028931632963940503, "epoch": 13.899295990205081, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2326096883.0, "reward": 8.014376497268676, "reward_std": 0.3895959258079529, "rewards/combined_reward/mean": 8.01437668800354, "rewards/combined_reward/std": 2.2077278792858124, "step": 22710 }, { "clip_ratio/high_max": 0.0022243103822499203, "clip_ratio/high_mean": 0.001312205241144208, "clip_ratio/low_mean": 0.0007726943995900608, "clip_ratio/low_min": 0.0001665325721357173, "clip_ratio/region_mean": 0.0020848996493441517, "completions/clipped_ratio": 0.0, "completions/max_length": 615.9, "completions/max_terminated_length": 615.9, "completions/mean_length": 201.8307357788086, "completions/mean_terminated_length": 201.8307357788086, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.026759145812441907, "epoch": 13.917661463116008, "frac_reward_zero_std": 0.18333333656191825, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 2329172073.0, "reward": 7.585907649993897, "reward_std": 0.40176817327737807, "rewards/combined_reward/mean": 7.585907363891602, "rewards/combined_reward/std": 2.7714893937110903, "step": 22740 }, { "clip_ratio/high_max": 0.0023377377423457802, "clip_ratio/high_mean": 0.001324647612515643, "clip_ratio/low_mean": 0.0009372635842737509, "clip_ratio/low_min": 0.00033144101544166913, "clip_ratio/region_mean": 0.0022619111953342025, "completions/clipped_ratio": 0.0, "completions/max_length": 558.4, "completions/max_terminated_length": 558.4, "completions/mean_length": 203.1632858276367, "completions/mean_terminated_length": 203.1632858276367, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.02980402355703215, "epoch": 13.936026936026936, "frac_reward_zero_std": 0.13333333507180214, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2332252380.0, "reward": 8.042477655410767, "reward_std": 0.3591073006391525, "rewards/combined_reward/mean": 8.042477416992188, "rewards/combined_reward/std": 2.218619680404663, "step": 22770 }, { "clip_ratio/high_max": 0.002216768602860005, "clip_ratio/high_mean": 0.0013507716021801267, "clip_ratio/low_mean": 0.0009175940411902654, "clip_ratio/low_min": 0.0002326900425638693, "clip_ratio/region_mean": 0.002268365646887105, "completions/clipped_ratio": 0.0, "completions/max_length": 645.8, "completions/max_terminated_length": 645.8, "completions/mean_length": 192.03307647705077, "completions/mean_terminated_length": 192.03307647705077, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02802377555053681, "epoch": 13.954392408937863, "frac_reward_zero_std": 0.17916667200624942, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2335289947.0, "reward": 7.897720527648926, "reward_std": 0.36247464567422866, "rewards/combined_reward/mean": 7.897720432281494, "rewards/combined_reward/std": 2.37312433719635, "step": 22800 }, { "clip_ratio/high_max": 0.0018874654289296208, "clip_ratio/high_mean": 0.0010877222630369943, "clip_ratio/low_mean": 0.0008799332585719337, "clip_ratio/low_min": 0.0002744897436059546, "clip_ratio/region_mean": 0.001967655517607151, "completions/clipped_ratio": 0.0, "completions/max_length": 591.9, "completions/max_terminated_length": 591.9, "completions/mean_length": 205.14219512939454, "completions/mean_terminated_length": 205.14219512939454, "completions/min_length": 49.1, "completions/min_terminated_length": 49.1, "entropy": 0.028604699966187277, "epoch": 13.972145699418427, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2338377853.0, "reward": 8.131264543533325, "reward_std": 0.3144479379057884, "rewards/combined_reward/mean": 8.131264400482177, "rewards/combined_reward/std": 2.1141562461853027, "step": 22830 }, { "clip_ratio/high_max": 0.0020235455614359427, "clip_ratio/high_mean": 0.001200140672638857, "clip_ratio/low_mean": 0.0007509109828485332, "clip_ratio/low_min": 0.00020313076917470121, "clip_ratio/region_mean": 0.001951051665431199, "completions/clipped_ratio": 0.0, "completions/max_length": 585.9, "completions/max_terminated_length": 585.9, "completions/mean_length": 200.87083892822267, "completions/mean_terminated_length": 200.87083892822267, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.027392008900642396, "epoch": 13.990511172329354, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.02783203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2341449357.0, "reward": 8.167792654037475, "reward_std": 0.3053123652935028, "rewards/combined_reward/mean": 8.16779260635376, "rewards/combined_reward/std": 2.0985581278800964, "step": 22860 }, { "clip_ratio/high_max": 0.0023412400685808026, "clip_ratio/high_mean": 0.0012294700599206824, "clip_ratio/low_mean": 0.0006413595610418449, "clip_ratio/low_min": 0.00013463741555317594, "clip_ratio/region_mean": 0.0018708296120166779, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0909090909091, "completions/max_terminated_length": 579.0909090909091, "completions/mean_length": 205.18111350319603, "completions/mean_terminated_length": 205.18111350319603, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02625352570970283, "epoch": 14.009182736455463, "frac_reward_zero_std": 0.17803030935200778, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2344846218.0, "reward": 8.04198442805897, "reward_std": 0.3544998290863904, "rewards/combined_reward/mean": 8.041984211314809, "rewards/combined_reward/std": 2.206461863084273, "step": 22890 }, { "clip_ratio/high_max": 0.002284600208804477, "clip_ratio/high_mean": 0.0013654745582850107, "clip_ratio/low_mean": 0.001042369288794968, "clip_ratio/low_min": 0.000325187883330121, "clip_ratio/region_mean": 0.002407843851445553, "completions/clipped_ratio": 0.0, "completions/max_length": 632.7, "completions/max_terminated_length": 632.7, "completions/mean_length": 205.43568115234376, "completions/mean_terminated_length": 205.43568115234376, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02795409515189628, "epoch": 14.02754820936639, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2347935251.0, "reward": 8.281083869934083, "reward_std": 0.2726853035390377, "rewards/combined_reward/mean": 8.281083679199218, "rewards/combined_reward/std": 2.136258363723755, "step": 22920 }, { "clip_ratio/high_max": 0.0021702951982054703, "clip_ratio/high_mean": 0.0012091481915073624, "clip_ratio/low_mean": 0.0008938757763341225, "clip_ratio/low_min": 0.0002554905059999631, "clip_ratio/region_mean": 0.0021030239930648047, "completions/clipped_ratio": 0.0, "completions/max_length": 537.2, "completions/max_terminated_length": 537.2, "completions/mean_length": 193.24974517822267, "completions/mean_terminated_length": 193.24974517822267, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.027777656035808226, "epoch": 14.045913682277318, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2350977490.0, "reward": 8.097196483612061, "reward_std": 0.38082931041717527, "rewards/combined_reward/mean": 8.09719614982605, "rewards/combined_reward/std": 2.1945496559143067, "step": 22950 }, { "clip_ratio/high_max": 0.0024361766440658054, "clip_ratio/high_mean": 0.001309363645486883, "clip_ratio/low_mean": 0.0008690749737676622, "clip_ratio/low_min": 0.0002328279944777023, "clip_ratio/region_mean": 0.002178438632108737, "completions/clipped_ratio": 0.0, "completions/max_length": 542.4, "completions/max_terminated_length": 542.4, "completions/mean_length": 198.69974517822266, "completions/mean_terminated_length": 198.69974517822266, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02695750610437244, "epoch": 14.064279155188245, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2354040657.0, "reward": 8.015284490585326, "reward_std": 0.33261004388332366, "rewards/combined_reward/mean": 8.015284109115601, "rewards/combined_reward/std": 2.3359602332115172, "step": 22980 }, { "clip_ratio/high_max": 0.001783908072199362, "clip_ratio/high_mean": 0.001058687085363393, "clip_ratio/low_mean": 0.0008530996314220829, "clip_ratio/low_min": 0.0002946048812494458, "clip_ratio/region_mean": 0.0019117867268505506, "completions/clipped_ratio": 0.0, "completions/max_length": 637.4, "completions/max_terminated_length": 637.4, "completions/mean_length": 205.87474670410157, "completions/mean_terminated_length": 205.87474670410157, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02556095295585692, "epoch": 14.082644628099173, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.004974365234375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2357131376.0, "reward": 8.192180967330932, "reward_std": 0.3150738120079041, "rewards/combined_reward/mean": 8.192181062698364, "rewards/combined_reward/std": 2.022401785850525, "step": 23010 }, { "clip_ratio/high_max": 0.0019151850943065559, "clip_ratio/high_mean": 0.0010926157910337983, "clip_ratio/low_mean": 0.0008399447474099968, "clip_ratio/low_min": 0.000216621691652108, "clip_ratio/region_mean": 0.0019325605360791087, "completions/clipped_ratio": 0.0, "completions/max_length": 649.8, "completions/max_terminated_length": 649.8, "completions/mean_length": 211.19896392822267, "completions/mean_terminated_length": 211.19896392822267, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02619429798796773, "epoch": 14.1010101010101, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.026123046875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2360242540.0, "reward": 8.090235376358033, "reward_std": 0.36075004637241365, "rewards/combined_reward/mean": 8.090235185623168, "rewards/combined_reward/std": 2.413428616523743, "step": 23040 }, { "clip_ratio/high_max": 0.002739408040361013, "clip_ratio/high_mean": 0.0014821953261465145, "clip_ratio/low_mean": 0.0009146327859828792, "clip_ratio/low_min": 0.0003317423931245382, "clip_ratio/region_mean": 0.00239682810594483, "completions/clipped_ratio": 0.0, "completions/max_length": 573.8, "completions/max_terminated_length": 573.8, "completions/mean_length": 206.43073272705078, "completions/mean_terminated_length": 206.43073272705078, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.033286899173011385, "epoch": 14.119375573921028, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.004974365234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2363335394.0, "reward": 8.104250144958495, "reward_std": 0.3294694319367409, "rewards/combined_reward/mean": 8.104249906539916, "rewards/combined_reward/std": 2.144959270954132, "step": 23070 }, { "clip_ratio/high_max": 0.00220857471722411, "clip_ratio/high_mean": 0.0012598806872119895, "clip_ratio/low_mean": 0.0008975311798470405, "clip_ratio/low_min": 0.00026873226549165946, "clip_ratio/region_mean": 0.0021574118824598068, "completions/clipped_ratio": 0.0, "completions/max_length": 571.8, "completions/max_terminated_length": 571.8, "completions/mean_length": 204.47005767822264, "completions/mean_terminated_length": 204.47005767822264, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02895950651727617, "epoch": 14.137741046831955, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.006561279296875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2366420719.0, "reward": 8.142032241821289, "reward_std": 0.30846755057573316, "rewards/combined_reward/mean": 8.142031860351562, "rewards/combined_reward/std": 2.229397988319397, "step": 23100 }, { "clip_ratio/high_max": 0.002282788607165761, "clip_ratio/high_mean": 0.0012716875434610605, "clip_ratio/low_mean": 0.001044082356626556, "clip_ratio/low_min": 0.0002845257234487993, "clip_ratio/region_mean": 0.002315769919853968, "completions/clipped_ratio": 0.0, "completions/max_length": 674.1, "completions/max_terminated_length": 674.1, "completions/mean_length": 204.07370147705078, "completions/mean_terminated_length": 204.07370147705078, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.030321573947245876, "epoch": 14.156106519742883, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0252685546875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2369504522.0, "reward": 8.129316234588623, "reward_std": 0.3192719295620918, "rewards/combined_reward/mean": 8.129316043853759, "rewards/combined_reward/std": 1.9764802217483521, "step": 23130 }, { "clip_ratio/high_max": 0.0020979318370033676, "clip_ratio/high_mean": 0.0011844021775080668, "clip_ratio/low_mean": 0.0007787888584061875, "clip_ratio/low_min": 0.00024036351812052695, "clip_ratio/region_mean": 0.001963191024939685, "completions/clipped_ratio": 0.0, "completions/max_length": 548.2, "completions/max_terminated_length": 548.2, "completions/mean_length": 209.2921936035156, "completions/mean_terminated_length": 209.2921936035156, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.02582064925227314, "epoch": 14.17447199265381, "frac_reward_zero_std": 0.19583334103226663, "grad_norm": 0.007354736328125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2372608364.0, "reward": 8.261509704589844, "reward_std": 0.2698610439896584, "rewards/combined_reward/mean": 8.261509704589844, "rewards/combined_reward/std": 2.1235966205596926, "step": 23160 }, { "clip_ratio/high_max": 0.0027327478123576536, "clip_ratio/high_mean": 0.0015596287603936313, "clip_ratio/low_mean": 0.0009802769190476587, "clip_ratio/low_min": 0.00021562801108908995, "clip_ratio/region_mean": 0.0025399056656169707, "completions/clipped_ratio": 0.0, "completions/max_length": 590.7, "completions/max_terminated_length": 590.7, "completions/mean_length": 194.1734405517578, "completions/mean_terminated_length": 194.1734405517578, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.030975739021475115, "epoch": 14.192837465564738, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2375654150.0, "reward": 8.10508484840393, "reward_std": 0.31093529760837557, "rewards/combined_reward/mean": 8.105084800720215, "rewards/combined_reward/std": 2.2954384088516235, "step": 23190 }, { "clip_ratio/high_max": 0.002163993665453745, "clip_ratio/high_mean": 0.001273210456686987, "clip_ratio/low_mean": 0.0009471284747633035, "clip_ratio/low_min": 0.0003123272414086387, "clip_ratio/region_mean": 0.0022203389220521786, "completions/clipped_ratio": 0.0, "completions/max_length": 568.3, "completions/max_terminated_length": 568.3, "completions/mean_length": 193.96589202880858, "completions/mean_terminated_length": 193.96589202880858, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027763535159950455, "epoch": 14.211202938475665, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2378699139.0, "reward": 8.175877809524536, "reward_std": 0.30570442602038383, "rewards/combined_reward/mean": 8.17587752342224, "rewards/combined_reward/std": 2.173522710800171, "step": 23220 }, { "clip_ratio/high_max": 0.002160712237188515, "clip_ratio/high_mean": 0.0012632377496629489, "clip_ratio/low_mean": 0.0009398662633126757, "clip_ratio/low_min": 0.0002867752918973565, "clip_ratio/region_mean": 0.002203104036137423, "completions/clipped_ratio": 0.0, "completions/max_length": 650.4, "completions/max_terminated_length": 650.4, "completions/mean_length": 203.18490142822264, "completions/mean_terminated_length": 203.18490142822264, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.02804392925463617, "epoch": 14.229568411386593, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2381779529.0, "reward": 8.023075294494628, "reward_std": 0.3322279304265976, "rewards/combined_reward/mean": 8.023075294494628, "rewards/combined_reward/std": 2.298227822780609, "step": 23250 }, { "clip_ratio/high_max": 0.0021033643555711024, "clip_ratio/high_mean": 0.001168179239660579, "clip_ratio/low_mean": 0.0008412004304166961, "clip_ratio/low_min": 0.0002349878595850896, "clip_ratio/region_mean": 0.0020093796754129772, "completions/clipped_ratio": 0.0, "completions/max_length": 527.9, "completions/max_terminated_length": 527.9, "completions/mean_length": 196.97865295410156, "completions/mean_terminated_length": 196.97865295410156, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.026520740357227623, "epoch": 14.24793388429752, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2384836087.0, "reward": 8.039327907562257, "reward_std": 0.2985458493232727, "rewards/combined_reward/mean": 8.03932752609253, "rewards/combined_reward/std": 2.2053639888763428, "step": 23280 }, { "clip_ratio/high_max": 0.0030676692093644913, "clip_ratio/high_mean": 0.0018234489176393253, "clip_ratio/low_mean": 0.0008481528308038833, "clip_ratio/low_min": 0.00027636251397780145, "clip_ratio/region_mean": 0.0026716017338912932, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 205.2838577270508, "completions/mean_terminated_length": 205.2838577270508, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03312175045721233, "epoch": 14.266299357208448, "frac_reward_zero_std": 0.14166666865348815, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2387924537.0, "reward": 8.158910417556763, "reward_std": 0.4207812249660492, "rewards/combined_reward/mean": 8.158909845352174, "rewards/combined_reward/std": 2.099813389778137, "step": 23310 }, { "clip_ratio/high_max": 0.0021894351346418263, "clip_ratio/high_mean": 0.0012057868790482948, "clip_ratio/low_mean": 0.0008196152642388673, "clip_ratio/low_min": 0.000169922795612365, "clip_ratio/region_mean": 0.0020254021418319706, "completions/clipped_ratio": 0.0, "completions/max_length": 604.5, "completions/max_terminated_length": 604.5, "completions/mean_length": 193.54193115234375, "completions/mean_terminated_length": 193.54193115234375, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026285299953694146, "epoch": 14.284664830119375, "frac_reward_zero_std": 0.21666667237877846, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2390967898.0, "reward": 8.002567577362061, "reward_std": 0.32918280363082886, "rewards/combined_reward/mean": 8.002567338943482, "rewards/combined_reward/std": 2.240756118297577, "step": 23340 }, { "clip_ratio/high_max": 0.0021973777499321536, "clip_ratio/high_mean": 0.0012807682277828766, "clip_ratio/low_mean": 0.0009674055355693175, "clip_ratio/low_min": 0.00023509791911540862, "clip_ratio/region_mean": 0.0022481737687485293, "completions/clipped_ratio": 0.0, "completions/max_length": 493.2, "completions/max_terminated_length": 493.2, "completions/mean_length": 195.37448425292968, "completions/mean_terminated_length": 195.37448425292968, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.02848308162453274, "epoch": 14.303030303030303, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2394018296.0, "reward": 8.026441287994384, "reward_std": 0.32371891736984254, "rewards/combined_reward/mean": 8.026441144943238, "rewards/combined_reward/std": 2.423492991924286, "step": 23370 }, { "clip_ratio/high_max": 0.0029131769163844488, "clip_ratio/high_mean": 0.001610627044647117, "clip_ratio/low_mean": 0.0010238696856201084, "clip_ratio/low_min": 0.0002199063269169225, "clip_ratio/region_mean": 0.0026344967307522895, "completions/clipped_ratio": 0.0, "completions/max_length": 656.4, "completions/max_terminated_length": 656.4, "completions/mean_length": 196.74271392822266, "completions/mean_terminated_length": 196.74271392822266, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.03702838500806441, "epoch": 14.32139577594123, "frac_reward_zero_std": 0.21250000596046448, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2397073948.0, "reward": 8.051063060760498, "reward_std": 0.3373041287064552, "rewards/combined_reward/mean": 8.051062726974488, "rewards/combined_reward/std": 2.310105931758881, "step": 23400 }, { "clip_ratio/high_max": 0.002306589565220444, "clip_ratio/high_mean": 0.0013322793281986378, "clip_ratio/low_mean": 0.0008928043528915927, "clip_ratio/low_min": 0.0002863743024742386, "clip_ratio/region_mean": 0.0022250836870322623, "completions/clipped_ratio": 0.0, "completions/max_length": 537.9, "completions/max_terminated_length": 537.9, "completions/mean_length": 188.2479248046875, "completions/mean_terminated_length": 188.2479248046875, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028623176210870346, "epoch": 14.339761248852158, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2400096980.0, "reward": 8.196807670593262, "reward_std": 0.3228308014571667, "rewards/combined_reward/mean": 8.196807479858398, "rewards/combined_reward/std": 2.040331554412842, "step": 23430 }, { "clip_ratio/high_max": 0.0024209419687394986, "clip_ratio/high_mean": 0.0013959546083545623, "clip_ratio/low_mean": 0.0011522210833694166, "clip_ratio/low_min": 0.0004085032531293109, "clip_ratio/region_mean": 0.002548175703365511, "completions/clipped_ratio": 0.0, "completions/max_length": 627.6, "completions/max_terminated_length": 627.6, "completions/mean_length": 209.76406707763672, "completions/mean_terminated_length": 209.76406707763672, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.030999476307382187, "epoch": 14.358126721763085, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.00885009765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2403202634.0, "reward": 8.149281549453736, "reward_std": 0.30557156205177305, "rewards/combined_reward/mean": 8.14928126335144, "rewards/combined_reward/std": 2.141436743736267, "step": 23460 }, { "clip_ratio/high_max": 0.002344680317522337, "clip_ratio/high_mean": 0.0012460511303894843, "clip_ratio/low_mean": 0.0009244191972053765, "clip_ratio/low_min": 0.00016689240680231403, "clip_ratio/region_mean": 0.0021704703317179034, "completions/clipped_ratio": 0.0, "completions/max_length": 539.6, "completions/max_terminated_length": 539.6, "completions/mean_length": 195.6695343017578, "completions/mean_terminated_length": 195.6695343017578, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02943125437013805, "epoch": 14.376492194674013, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.00701904296875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2406254165.0, "reward": 8.300131559371948, "reward_std": 0.2694113448262215, "rewards/combined_reward/mean": 8.300131130218507, "rewards/combined_reward/std": 2.0635255932807923, "step": 23490 }, { "clip_ratio/high_max": 0.0021952235023491085, "clip_ratio/high_mean": 0.0012338726429637367, "clip_ratio/low_mean": 0.0007881478083315111, "clip_ratio/low_min": 0.0001629620642537096, "clip_ratio/region_mean": 0.0020220204542662637, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 194.6940155029297, "completions/mean_terminated_length": 194.6940155029297, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.026384768184895317, "epoch": 14.39485766758494, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.02001953125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2409301950.0, "reward": 8.36974573135376, "reward_std": 0.27447078824043275, "rewards/combined_reward/mean": 8.369745540618897, "rewards/combined_reward/std": 1.902932333946228, "step": 23520 }, { "clip_ratio/high_max": 0.002424726874839204, "clip_ratio/high_mean": 0.0014244743356054339, "clip_ratio/low_mean": 0.0009095914439967601, "clip_ratio/low_min": 0.0002737841711981067, "clip_ratio/region_mean": 0.002334065734369991, "completions/clipped_ratio": 0.0, "completions/max_length": 623.8, "completions/max_terminated_length": 623.8, "completions/mean_length": 190.81250457763673, "completions/mean_terminated_length": 190.81250457763673, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.027761803179358443, "epoch": 14.413223140495868, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2412334830.0, "reward": 7.9696752548217775, "reward_std": 0.2850410774350166, "rewards/combined_reward/mean": 7.969674873352051, "rewards/combined_reward/std": 2.254699194431305, "step": 23550 }, { "clip_ratio/high_max": 0.0025317927851574496, "clip_ratio/high_mean": 0.0013980875782484267, "clip_ratio/low_mean": 0.0008590453425616336, "clip_ratio/low_min": 0.00023642808131019895, "clip_ratio/region_mean": 0.0022571329138979007, "completions/clipped_ratio": 0.0, "completions/max_length": 548.2, "completions/max_terminated_length": 548.2, "completions/mean_length": 200.15364990234374, "completions/mean_terminated_length": 200.15364990234374, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030975169866966704, "epoch": 14.431588613406795, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 2415403580.0, "reward": 8.041015434265137, "reward_std": 0.31322375014424325, "rewards/combined_reward/mean": 8.041015195846558, "rewards/combined_reward/std": 2.4503170609474183, "step": 23580 }, { "clip_ratio/high_max": 0.002054054274049122, "clip_ratio/high_mean": 0.0012438973905470145, "clip_ratio/low_mean": 0.0009338795880466932, "clip_ratio/low_min": 0.0002274326849146746, "clip_ratio/region_mean": 0.002177776969134963, "completions/clipped_ratio": 0.0, "completions/max_length": 531.8, "completions/max_terminated_length": 531.8, "completions/mean_length": 192.24193420410157, "completions/mean_terminated_length": 192.24193420410157, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.027355937869288027, "epoch": 14.449954086317723, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2418441949.0, "reward": 7.8766350746154785, "reward_std": 0.30242799073457716, "rewards/combined_reward/mean": 7.876634788513184, "rewards/combined_reward/std": 2.4305931329727173, "step": 23610 }, { "clip_ratio/high_max": 0.0025658538230345584, "clip_ratio/high_mean": 0.0013558370539006623, "clip_ratio/low_mean": 0.0008295683364849538, "clip_ratio/low_min": 0.0002604213940988605, "clip_ratio/region_mean": 0.0021854053877177647, "completions/clipped_ratio": 0.0, "completions/max_length": 558.6, "completions/max_terminated_length": 558.6, "completions/mean_length": 204.82265930175782, "completions/mean_terminated_length": 204.82265930175782, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.028294087604930004, "epoch": 14.46831955922865, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.0296630859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2421528628.0, "reward": 8.185326290130615, "reward_std": 0.2979889243841171, "rewards/combined_reward/mean": 8.185325813293456, "rewards/combined_reward/std": 2.264487564563751, "step": 23640 }, { "clip_ratio/high_max": 0.0024993208411615343, "clip_ratio/high_mean": 0.0013698305305297255, "clip_ratio/low_mean": 0.0009162593150297955, "clip_ratio/low_min": 0.0002395704374066554, "clip_ratio/region_mean": 0.0022860898689638512, "completions/clipped_ratio": 0.0, "completions/max_length": 517.6, "completions/max_terminated_length": 517.6, "completions/mean_length": 186.33333892822264, "completions/mean_terminated_length": 186.33333892822264, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029085591761395337, "epoch": 14.486685032139578, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.00732421875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 2424544308.0, "reward": 8.138213777542115, "reward_std": 0.2931228011846542, "rewards/combined_reward/mean": 8.138213777542115, "rewards/combined_reward/std": 2.343119764328003, "step": 23670 }, { "clip_ratio/high_max": 0.002315404587231266, "clip_ratio/high_mean": 0.0012730655899455693, "clip_ratio/low_mean": 0.0006748801803041715, "clip_ratio/low_min": 0.00014757384851691314, "clip_ratio/region_mean": 0.0019479457643077089, "completions/clipped_ratio": 0.0, "completions/max_length": 706.3, "completions/max_terminated_length": 706.3, "completions/mean_length": 195.7726623535156, "completions/mean_terminated_length": 195.7726623535156, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.024558747901270786, "epoch": 14.505050505050505, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.006317138671875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2427596235.0, "reward": 7.96990065574646, "reward_std": 0.33643939197063444, "rewards/combined_reward/mean": 7.969900417327881, "rewards/combined_reward/std": 2.316468322277069, "step": 23700 }, { "clip_ratio/high_max": 0.003005270656043043, "clip_ratio/high_mean": 0.00182539540498207, "clip_ratio/low_mean": 0.001167686843594614, "clip_ratio/low_min": 0.00023980996144625047, "clip_ratio/region_mean": 0.002993082233782237, "completions/clipped_ratio": 0.0, "completions/max_length": 620.1, "completions/max_terminated_length": 620.1, "completions/mean_length": 196.28958892822266, "completions/mean_terminated_length": 196.28958892822266, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.03647421368708213, "epoch": 14.523415977961433, "frac_reward_zero_std": 0.19166667200624943, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2430650147.0, "reward": 7.888670587539673, "reward_std": 0.2658233255147934, "rewards/combined_reward/mean": 7.88867039680481, "rewards/combined_reward/std": 2.5363429307937624, "step": 23730 }, { "clip_ratio/high_max": 0.0019005138979991897, "clip_ratio/high_mean": 0.0010976601222258372, "clip_ratio/low_mean": 0.000832631489902269, "clip_ratio/low_min": 0.0002497015031015811, "clip_ratio/region_mean": 0.0019302915995164465, "completions/clipped_ratio": 0.0, "completions/max_length": 590.8, "completions/max_terminated_length": 590.8, "completions/mean_length": 202.9924530029297, "completions/mean_terminated_length": 202.9924530029297, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.026131950792235632, "epoch": 14.54178145087236, "frac_reward_zero_std": 0.1833333346992731, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2433729798.0, "reward": 8.393566226959228, "reward_std": 0.3203038215637207, "rewards/combined_reward/mean": 8.39356575012207, "rewards/combined_reward/std": 1.8666630625724792, "step": 23760 }, { "clip_ratio/high_max": 0.002950700389434739, "clip_ratio/high_mean": 0.0016470792254646464, "clip_ratio/low_mean": 0.000998452410082488, "clip_ratio/low_min": 0.0003204855755029712, "clip_ratio/region_mean": 0.0026455316265734533, "completions/clipped_ratio": 0.0, "completions/max_length": 641.8, "completions/max_terminated_length": 641.8, "completions/mean_length": 200.73906860351562, "completions/mean_terminated_length": 200.73906860351562, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.029159488839407763, "epoch": 14.560146923783288, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2436800796.0, "reward": 7.991835260391236, "reward_std": 0.386263045668602, "rewards/combined_reward/mean": 7.99183497428894, "rewards/combined_reward/std": 2.312612223625183, "step": 23790 }, { "clip_ratio/high_max": 0.0022985680795197063, "clip_ratio/high_mean": 0.0013662548017843315, "clip_ratio/low_mean": 0.0011390369031384277, "clip_ratio/low_min": 0.0004076043493114412, "clip_ratio/region_mean": 0.0025052917073480783, "completions/clipped_ratio": 0.0, "completions/max_length": 617.9, "completions/max_terminated_length": 617.9, "completions/mean_length": 206.00443267822266, "completions/mean_terminated_length": 206.00443267822266, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.02944663125090301, "epoch": 14.578512396694215, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2439892013.0, "reward": 7.982307481765747, "reward_std": 0.34549793154001235, "rewards/combined_reward/mean": 7.9823075294494625, "rewards/combined_reward/std": 2.207359755039215, "step": 23820 }, { "clip_ratio/high_max": 0.002002859514565595, "clip_ratio/high_mean": 0.001138050132067292, "clip_ratio/low_mean": 0.0008796996153250802, "clip_ratio/low_min": 0.0002647895007006203, "clip_ratio/region_mean": 0.0020177497404802124, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 200.70391082763672, "completions/mean_terminated_length": 200.70391082763672, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02652485407112787, "epoch": 14.596877869605143, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2442962876.0, "reward": 8.478780269622803, "reward_std": 0.29803196936845777, "rewards/combined_reward/mean": 8.478780221939086, "rewards/combined_reward/std": 1.8953482151031493, "step": 23850 }, { "clip_ratio/high_max": 0.002291443360930619, "clip_ratio/high_mean": 0.0013565709476097254, "clip_ratio/low_mean": 0.0008545972067319478, "clip_ratio/low_min": 0.00023324710806870523, "clip_ratio/region_mean": 0.002211168175684482, "completions/clipped_ratio": 0.0, "completions/max_length": 563.1, "completions/max_terminated_length": 563.1, "completions/mean_length": 199.05026550292968, "completions/mean_terminated_length": 199.05026550292968, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028169272102726003, "epoch": 14.61524334251607, "frac_reward_zero_std": 0.17500000335276128, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2446027389.0, "reward": 8.09024682044983, "reward_std": 0.2961890235543251, "rewards/combined_reward/mean": 8.090246438980103, "rewards/combined_reward/std": 2.189856004714966, "step": 23880 }, { "clip_ratio/high_max": 0.0025133525026224865, "clip_ratio/high_mean": 0.0013904090661526424, "clip_ratio/low_mean": 0.0011881474779026273, "clip_ratio/low_min": 0.00035181082760876357, "clip_ratio/region_mean": 0.002578556530109684, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 191.54870147705077, "completions/mean_terminated_length": 191.54870147705077, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03264603412244469, "epoch": 14.633608815426998, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 2449063096.0, "reward": 7.837694501876831, "reward_std": 0.3291626751422882, "rewards/combined_reward/mean": 7.837694406509399, "rewards/combined_reward/std": 2.268964374065399, "step": 23910 }, { "clip_ratio/high_max": 0.0023248619506678852, "clip_ratio/high_mean": 0.0013006644066384373, "clip_ratio/low_mean": 0.0009914575748553034, "clip_ratio/low_min": 0.0002531027637208657, "clip_ratio/region_mean": 0.00229212199531806, "completions/clipped_ratio": 0.0, "completions/max_length": 671.2, "completions/max_terminated_length": 671.2, "completions/mean_length": 205.9674514770508, "completions/mean_terminated_length": 205.9674514770508, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.028723367971057692, "epoch": 14.651974288337925, "frac_reward_zero_std": 0.1333333369344473, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2452154171.0, "reward": 7.999360132217407, "reward_std": 0.32074377685785294, "rewards/combined_reward/mean": 7.999359607696533, "rewards/combined_reward/std": 2.2788611888885497, "step": 23940 }, { "clip_ratio/high_max": 0.0022052930185357885, "clip_ratio/high_mean": 0.0012381941446316584, "clip_ratio/low_mean": 0.0009220330714015291, "clip_ratio/low_min": 0.00018076598886788511, "clip_ratio/region_mean": 0.0021602272288873793, "completions/clipped_ratio": 0.0, "completions/max_length": 569.4, "completions/max_terminated_length": 569.4, "completions/mean_length": 193.59427337646486, "completions/mean_terminated_length": 193.59427337646486, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.0277827228885144, "epoch": 14.670339761248853, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2455197733.0, "reward": 8.185761213302612, "reward_std": 0.28383664786815643, "rewards/combined_reward/mean": 8.185761165618896, "rewards/combined_reward/std": 2.094661796092987, "step": 23970 }, { "clip_ratio/high_max": 0.0025654829606840696, "clip_ratio/high_mean": 0.0014419843223246668, "clip_ratio/low_mean": 0.0009339622557793821, "clip_ratio/low_min": 0.0002477847650879994, "clip_ratio/region_mean": 0.0023759465611268146, "completions/clipped_ratio": 0.0, "completions/max_length": 692.1, "completions/max_terminated_length": 692.1, "completions/mean_length": 213.6710998535156, "completions/mean_terminated_length": 213.6710998535156, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.029102462468047936, "epoch": 14.68870523415978, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.027099609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2458318390.0, "reward": 7.888222694396973, "reward_std": 0.37979167997837066, "rewards/combined_reward/mean": 7.8882224559783936, "rewards/combined_reward/std": 2.611780047416687, "step": 24000 }, { "clip_ratio/high_max": 0.0028155492007499562, "clip_ratio/high_mean": 0.0015230850525161561, "clip_ratio/low_mean": 0.0011761360659268878, "clip_ratio/low_min": 0.0003551561695834001, "clip_ratio/region_mean": 0.002699221106013283, "completions/clipped_ratio": 0.0, "completions/max_length": 659.5, "completions/max_terminated_length": 659.5, "completions/mean_length": 200.32604522705077, "completions/mean_terminated_length": 200.32604522705077, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029769352249180276, "epoch": 14.707070707070708, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2461387802.0, "reward": 7.899860143661499, "reward_std": 0.3351084887981415, "rewards/combined_reward/mean": 7.899859952926636, "rewards/combined_reward/std": 2.349741756916046, "step": 24030 }, { "clip_ratio/high_max": 0.0023900075543982287, "clip_ratio/high_mean": 0.0014709711757556457, "clip_ratio/low_mean": 0.0009552281512393772, "clip_ratio/low_min": 0.00019342683323581394, "clip_ratio/region_mean": 0.002426199327843885, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 207.36562805175782, "completions/mean_terminated_length": 207.36562805175782, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.030257737527911863, "epoch": 14.725436179981635, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2464484246.0, "reward": 8.118948936462402, "reward_std": 0.3000522285699844, "rewards/combined_reward/mean": 8.118948841094971, "rewards/combined_reward/std": 2.405942678451538, "step": 24060 }, { "clip_ratio/high_max": 0.0024889026783057487, "clip_ratio/high_mean": 0.00149165954523293, "clip_ratio/low_mean": 0.0010020703293169694, "clip_ratio/low_min": 0.0002727269978398302, "clip_ratio/region_mean": 0.0024937298798856015, "completions/clipped_ratio": 0.0, "completions/max_length": 540.3, "completions/max_terminated_length": 540.3, "completions/mean_length": 199.18073272705078, "completions/mean_terminated_length": 199.18073272705078, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02893328663582603, "epoch": 14.743801652892563, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2467549260.0, "reward": 7.919698476791382, "reward_std": 0.2853868529200554, "rewards/combined_reward/mean": 7.919698429107666, "rewards/combined_reward/std": 2.4154126405715943, "step": 24090 }, { "clip_ratio/high_max": 0.0026982224313542246, "clip_ratio/high_mean": 0.0016011457123871272, "clip_ratio/low_mean": 0.0009920616310409968, "clip_ratio/low_min": 0.0002525624731788412, "clip_ratio/region_mean": 0.0025932073360309003, "completions/clipped_ratio": 0.0, "completions/max_length": 599.9, "completions/max_terminated_length": 599.9, "completions/mean_length": 206.96771392822265, "completions/mean_terminated_length": 206.96771392822265, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.032401956943795086, "epoch": 14.76216712580349, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2470644176.0, "reward": 7.895067167282105, "reward_std": 0.3689309537410736, "rewards/combined_reward/mean": 7.895066738128662, "rewards/combined_reward/std": 2.337584912776947, "step": 24120 }, { "clip_ratio/high_max": 0.0028255862421550167, "clip_ratio/high_mean": 0.0017100176529008118, "clip_ratio/low_mean": 0.001186141224267582, "clip_ratio/low_min": 0.00038869387450783205, "clip_ratio/region_mean": 0.0028961588754706706, "completions/clipped_ratio": 0.0, "completions/max_length": 561.7, "completions/max_terminated_length": 561.7, "completions/mean_length": 205.8635452270508, "completions/mean_terminated_length": 205.8635452270508, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.035367025341838595, "epoch": 14.780532598714418, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2473734852.0, "reward": 7.815819168090821, "reward_std": 0.37237253338098525, "rewards/combined_reward/mean": 7.815818881988525, "rewards/combined_reward/std": 2.465402293205261, "step": 24150 }, { "clip_ratio/high_max": 0.0023496158833343846, "clip_ratio/high_mean": 0.0012982185024156933, "clip_ratio/low_mean": 0.0010920867699799905, "clip_ratio/low_min": 0.00023005060502327978, "clip_ratio/region_mean": 0.0023903052671812473, "completions/clipped_ratio": 0.0, "completions/max_length": 609.7, "completions/max_terminated_length": 609.7, "completions/mean_length": 191.3940170288086, "completions/mean_terminated_length": 191.3940170288086, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.030990218045189977, "epoch": 14.798898071625345, "frac_reward_zero_std": 0.19166667200624943, "grad_norm": 0.04150390625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2476769965.0, "reward": 7.809561777114868, "reward_std": 0.34410167932510377, "rewards/combined_reward/mean": 7.809561634063721, "rewards/combined_reward/std": 2.3142096042633056, "step": 24180 }, { "clip_ratio/high_max": 0.002480672520323424, "clip_ratio/high_mean": 0.001433876497564294, "clip_ratio/low_mean": 0.0010963499383554638, "clip_ratio/low_min": 0.00022901795479507808, "clip_ratio/region_mean": 0.002530226419912651, "completions/clipped_ratio": 0.0, "completions/max_length": 478.4, "completions/max_terminated_length": 478.4, "completions/mean_length": 195.27032012939452, "completions/mean_terminated_length": 195.27032012939452, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.032076967088505624, "epoch": 14.817263544536273, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 2479819963.0, "reward": 8.290901851654052, "reward_std": 0.2824196919798851, "rewards/combined_reward/mean": 8.290901899337769, "rewards/combined_reward/std": 2.0099861145019533, "step": 24210 }, { "clip_ratio/high_max": 0.0023045628578984177, "clip_ratio/high_mean": 0.0013301299111238526, "clip_ratio/low_mean": 0.001150136295837001, "clip_ratio/low_min": 0.0003724479400261771, "clip_ratio/region_mean": 0.002480266190347417, "completions/clipped_ratio": 0.0, "completions/max_length": 563.9, "completions/max_terminated_length": 563.9, "completions/mean_length": 204.19271545410157, "completions/mean_terminated_length": 204.19271545410157, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.030351452032725015, "epoch": 14.8356290174472, "frac_reward_zero_std": 0.2000000048428774, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2482904223.0, "reward": 7.787011337280274, "reward_std": 0.3774204567074776, "rewards/combined_reward/mean": 7.787011241912841, "rewards/combined_reward/std": 2.43079559803009, "step": 24240 }, { "clip_ratio/high_max": 0.0022699012651476854, "clip_ratio/high_mean": 0.0012612751491663705, "clip_ratio/low_mean": 0.0009169472761035043, "clip_ratio/low_min": 0.00027309539606600687, "clip_ratio/region_mean": 0.002178222418782146, "completions/clipped_ratio": 0.0, "completions/max_length": 570.8, "completions/max_terminated_length": 570.8, "completions/mean_length": 196.6901107788086, "completions/mean_terminated_length": 196.6901107788086, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02784117479653408, "epoch": 14.853994490358126, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2485959673.0, "reward": 7.953005456924439, "reward_std": 0.34009767323732376, "rewards/combined_reward/mean": 7.953005456924439, "rewards/combined_reward/std": 2.195588231086731, "step": 24270 }, { "clip_ratio/high_max": 0.0027268470148555934, "clip_ratio/high_mean": 0.0015884082445457655, "clip_ratio/low_mean": 0.001063889709136371, "clip_ratio/low_min": 0.0002986074658110738, "clip_ratio/region_mean": 0.0026522979630196155, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 191.26901550292968, "completions/mean_terminated_length": 191.26901550292968, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03292615630974372, "epoch": 14.872359963269055, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2488994306.0, "reward": 7.950176954269409, "reward_std": 0.35889213979244233, "rewards/combined_reward/mean": 7.950176668167114, "rewards/combined_reward/std": 2.4238025903701783, "step": 24300 }, { "clip_ratio/high_max": 0.002196036789488668, "clip_ratio/high_mean": 0.0012990673434008688, "clip_ratio/low_mean": 0.0010259433339039484, "clip_ratio/low_min": 0.0003687751586160933, "clip_ratio/region_mean": 0.0023250106857934344, "completions/clipped_ratio": 0.0, "completions/max_length": 593.4, "completions/max_terminated_length": 593.4, "completions/mean_length": 202.21328735351562, "completions/mean_terminated_length": 202.21328735351562, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.029785456617052355, "epoch": 14.89072543617998, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2492070965.0, "reward": 7.779020690917969, "reward_std": 0.34349341988563536, "rewards/combined_reward/mean": 7.7790204048156735, "rewards/combined_reward/std": 2.3139335751533507, "step": 24330 }, { "clip_ratio/high_max": 0.0021001168148359283, "clip_ratio/high_mean": 0.0011799733754742192, "clip_ratio/low_mean": 0.0010415933516924269, "clip_ratio/low_min": 0.00027465948417860396, "clip_ratio/region_mean": 0.002221566731653487, "completions/clipped_ratio": 0.0, "completions/max_length": 604.4, "completions/max_terminated_length": 604.4, "completions/mean_length": 203.9705780029297, "completions/mean_terminated_length": 203.9705780029297, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02916355198249221, "epoch": 14.909090909090908, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 2495154372.0, "reward": 7.876371812820435, "reward_std": 0.3734911471605301, "rewards/combined_reward/mean": 7.87637152671814, "rewards/combined_reward/std": 2.4619715332984926, "step": 24360 }, { "clip_ratio/high_max": 0.0021022150855666646, "clip_ratio/high_mean": 0.0011978039547102526, "clip_ratio/low_mean": 0.0009701021172077162, "clip_ratio/low_min": 0.000161066630971618, "clip_ratio/region_mean": 0.0021679060577298515, "completions/clipped_ratio": 0.0, "completions/max_length": 584.3, "completions/max_terminated_length": 584.3, "completions/mean_length": 197.45365142822266, "completions/mean_terminated_length": 197.45365142822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.029001054889522492, "epoch": 14.927456382001836, "frac_reward_zero_std": 0.19583333916962148, "grad_norm": 0.027099609375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2498212754.0, "reward": 7.882753181457519, "reward_std": 0.3997239112854004, "rewards/combined_reward/mean": 7.882753086090088, "rewards/combined_reward/std": 2.3324729800224304, "step": 24390 }, { "clip_ratio/high_max": 0.00289494590688264, "clip_ratio/high_mean": 0.0016355715764802881, "clip_ratio/low_mean": 0.0007906263633534157, "clip_ratio/low_min": 0.00019194693741155788, "clip_ratio/region_mean": 0.0024261979279496396, "completions/clipped_ratio": 0.0, "completions/max_length": 783.2, "completions/max_terminated_length": 783.2, "completions/mean_length": 199.1362045288086, "completions/mean_terminated_length": 199.1362045288086, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.03293412172545989, "epoch": 14.945821854912763, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.0130615234375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2501277597.0, "reward": 7.744356966018676, "reward_std": 0.3472678601741791, "rewards/combined_reward/mean": 7.744356775283814, "rewards/combined_reward/std": 2.646364998817444, "step": 24420 }, { "clip_ratio/high_max": 0.0027012506465932044, "clip_ratio/high_mean": 0.001720477380392064, "clip_ratio/low_mean": 0.0010150613510631957, "clip_ratio/low_min": 0.0003424241952113031, "clip_ratio/region_mean": 0.002735538738488685, "completions/clipped_ratio": 0.0, "completions/max_length": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 193.3101608276367, "completions/mean_terminated_length": 193.3101608276367, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.03321652067825198, "epoch": 14.96418732782369, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2504320068.0, "reward": 8.125482177734375, "reward_std": 0.403232553601265, "rewards/combined_reward/mean": 8.125481939315796, "rewards/combined_reward/std": 2.212031590938568, "step": 24450 }, { "clip_ratio/high_max": 0.0022735038733420272, "clip_ratio/high_mean": 0.001249630539920569, "clip_ratio/low_mean": 0.0008953191125328885, "clip_ratio/low_min": 0.0001750112166822267, "clip_ratio/region_mean": 0.00214494964457117, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 204.56328430175782, "completions/mean_terminated_length": 204.56328430175782, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.02844733321107924, "epoch": 14.982552800734618, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2507405751.0, "reward": 8.046165657043456, "reward_std": 0.35255433320999147, "rewards/combined_reward/mean": 8.04616551399231, "rewards/combined_reward/std": 2.2034010529518127, "step": 24480 }, { "clip_ratio/high_max": 0.0024433093970755936, "clip_ratio/high_mean": 0.0014159404254623026, "clip_ratio/low_mean": 0.0010810580511056518, "clip_ratio/low_min": 0.0002745893476550641, "clip_ratio/region_mean": 0.0024969984902566203, "completions/clipped_ratio": 0.0, "completions/max_length": 808.5555555555555, "completions/max_terminated_length": 808.5555555555555, "completions/mean_length": 186.53501553005643, "completions/mean_terminated_length": 186.53501553005643, "completions/min_length": 37.22222222222222, "completions/min_terminated_length": 37.22222222222222, "entropy": 0.03186624225671009, "epoch": 15.000612182430364, "frac_reward_zero_std": 0.1944444498254193, "grad_norm": 0.00543212890625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2510120560.0, "reward": 8.097662448883057, "reward_std": 0.29467441472742295, "rewards/combined_reward/mean": 8.097662130991617, "rewards/combined_reward/std": 2.1319495836893716, "step": 24510 }, { "clip_ratio/high_max": 0.0020033101609442383, "clip_ratio/high_mean": 0.0010677944306128969, "clip_ratio/low_mean": 0.0006942394486638174, "clip_ratio/low_min": 0.00020541819831123575, "clip_ratio/region_mean": 0.0017620338641184694, "completions/clipped_ratio": 0.0, "completions/max_length": 615.8, "completions/max_terminated_length": 615.8, "completions/mean_length": 196.05469360351563, "completions/mean_terminated_length": 196.05469360351563, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.028932023234665392, "epoch": 15.018977655341292, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.007171630859375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2513173570.0, "reward": 7.952254915237427, "reward_std": 0.2851637423038483, "rewards/combined_reward/mean": 7.952254724502564, "rewards/combined_reward/std": 2.285819780826569, "step": 24540 }, { "clip_ratio/high_max": 0.001891685176815372, "clip_ratio/high_mean": 0.0010459975851214646, "clip_ratio/low_mean": 0.0006420989312876676, "clip_ratio/low_min": 0.0002096722659189254, "clip_ratio/region_mean": 0.0016880965258072441, "completions/clipped_ratio": 0.0, "completions/max_length": 549.1, "completions/max_terminated_length": 549.1, "completions/mean_length": 198.84583892822266, "completions/mean_terminated_length": 198.84583892822266, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02847391654116412, "epoch": 15.03734312825222, "frac_reward_zero_std": 0.11250000335276127, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2516237298.0, "reward": 7.973417472839356, "reward_std": 0.3867593422532082, "rewards/combined_reward/mean": 7.9734172344207765, "rewards/combined_reward/std": 2.3184168100357057, "step": 24570 }, { "clip_ratio/high_max": 0.001982446975792603, "clip_ratio/high_mean": 0.0011287210288476975, "clip_ratio/low_mean": 0.0008298933508437283, "clip_ratio/low_min": 0.00023587210259089867, "clip_ratio/region_mean": 0.001958614363199255, "completions/clipped_ratio": 0.0, "completions/max_length": 596.6, "completions/max_terminated_length": 596.6, "completions/mean_length": 192.1838592529297, "completions/mean_terminated_length": 192.1838592529297, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.02993328825881084, "epoch": 15.055708601163147, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01434326171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2519275444.0, "reward": 8.002155637741089, "reward_std": 0.36568291783332824, "rewards/combined_reward/mean": 8.002155303955078, "rewards/combined_reward/std": 2.3690782308578493, "step": 24600 }, { "clip_ratio/high_max": 0.0020277267181275723, "clip_ratio/high_mean": 0.0010709337477843898, "clip_ratio/low_mean": 0.000695468519249213, "clip_ratio/low_min": 0.0002316529724339489, "clip_ratio/region_mean": 0.0017664022617585337, "completions/clipped_ratio": 0.0, "completions/max_length": 579.1, "completions/max_terminated_length": 579.1, "completions/mean_length": 182.94583892822266, "completions/mean_terminated_length": 182.94583892822266, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.029896585294045507, "epoch": 15.074074074074074, "frac_reward_zero_std": 0.2041666701436043, "grad_norm": 0.021728515625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2522278116.0, "reward": 8.233420181274415, "reward_std": 0.270308093726635, "rewards/combined_reward/mean": 8.233419942855836, "rewards/combined_reward/std": 2.117738127708435, "step": 24630 }, { "clip_ratio/high_max": 0.0016461914676862458, "clip_ratio/high_mean": 0.0009528264936913426, "clip_ratio/low_mean": 0.0005740434464920933, "clip_ratio/low_min": 0.00017307170128333382, "clip_ratio/region_mean": 0.001526869946004202, "completions/clipped_ratio": 0.0, "completions/max_length": 617.2, "completions/max_terminated_length": 617.2, "completions/mean_length": 190.76276397705078, "completions/mean_terminated_length": 190.76276397705078, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.024590589463089903, "epoch": 15.092439546985002, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2525310805.0, "reward": 8.23403091430664, "reward_std": 0.28309897929430006, "rewards/combined_reward/mean": 8.234030437469482, "rewards/combined_reward/std": 2.0968485951423643, "step": 24660 }, { "clip_ratio/high_max": 0.0019242187913429613, "clip_ratio/high_mean": 0.0010872715278916682, "clip_ratio/low_mean": 0.0007524408272729488, "clip_ratio/low_min": 0.0001872876649334406, "clip_ratio/region_mean": 0.0018397123688676704, "completions/clipped_ratio": 0.0, "completions/max_length": 571.7, "completions/max_terminated_length": 571.7, "completions/mean_length": 199.00443267822266, "completions/mean_terminated_length": 199.00443267822266, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027206373882169524, "epoch": 15.11080501989593, "frac_reward_zero_std": 0.12500000260770322, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2528375142.0, "reward": 8.151368618011475, "reward_std": 0.3162464663386345, "rewards/combined_reward/mean": 8.151368284225464, "rewards/combined_reward/std": 2.18495135307312, "step": 24690 }, { "clip_ratio/high_max": 0.0019609758945686433, "clip_ratio/high_mean": 0.0012011840677587315, "clip_ratio/low_mean": 0.0006779346231875631, "clip_ratio/low_min": 0.00017322584923628407, "clip_ratio/region_mean": 0.0018791187008901033, "completions/clipped_ratio": 0.0, "completions/max_length": 533.1, "completions/max_terminated_length": 533.1, "completions/mean_length": 201.82760925292968, "completions/mean_terminated_length": 201.82760925292968, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.02740519978882124, "epoch": 15.129170492806857, "frac_reward_zero_std": 0.22083333805203437, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2531450320.0, "reward": 8.054643917083741, "reward_std": 0.3252596765756607, "rewards/combined_reward/mean": 8.054643678665162, "rewards/combined_reward/std": 2.419524383544922, "step": 24720 }, { "clip_ratio/high_max": 0.002169956963431711, "clip_ratio/high_mean": 0.0011679945530583306, "clip_ratio/low_mean": 0.0006675083728623577, "clip_ratio/low_min": 0.00016427602519494637, "clip_ratio/region_mean": 0.0018355029334391779, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 196.05052642822267, "completions/mean_terminated_length": 196.05052642822267, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03075471051658193, "epoch": 15.147535965717784, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.005279541015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2534503314.0, "reward": 7.96448245048523, "reward_std": 0.3372479289770126, "rewards/combined_reward/mean": 7.964482259750366, "rewards/combined_reward/std": 2.3049301743507384, "step": 24750 }, { "clip_ratio/high_max": 0.002090168332021373, "clip_ratio/high_mean": 0.001225716468131092, "clip_ratio/low_mean": 0.0007442391712781197, "clip_ratio/low_min": 0.00020509622190729716, "clip_ratio/region_mean": 0.0019699556491104887, "completions/clipped_ratio": 0.0, "completions/max_length": 572.8, "completions/max_terminated_length": 572.8, "completions/mean_length": 215.53620452880858, "completions/mean_terminated_length": 215.53620452880858, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.032832198900481065, "epoch": 15.165901438628712, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2537631133.0, "reward": 8.108458185195923, "reward_std": 0.368183408677578, "rewards/combined_reward/mean": 8.108457851409913, "rewards/combined_reward/std": 2.2660547733306884, "step": 24780 }, { "clip_ratio/high_max": 0.0016618297711829654, "clip_ratio/high_mean": 0.000966379526653327, "clip_ratio/low_mean": 0.0006512996424135054, "clip_ratio/low_min": 0.000239141008205479, "clip_ratio/region_mean": 0.001617679189075716, "completions/clipped_ratio": 0.0, "completions/max_length": 499.5, "completions/max_terminated_length": 499.5, "completions/mean_length": 183.0670623779297, "completions/mean_terminated_length": 183.0670623779297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029686006158590315, "epoch": 15.183654729109275, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2540993904.0, "reward": 7.838291168212891, "reward_std": 0.36070771142840385, "rewards/combined_reward/mean": 7.838290810585022, "rewards/combined_reward/std": 2.463946968317032, "step": 24810 }, { "clip_ratio/high_max": 0.0018694437496984998, "clip_ratio/high_mean": 0.0010551036715090352, "clip_ratio/low_mean": 0.0008180290379338355, "clip_ratio/low_min": 0.00025566052693951253, "clip_ratio/region_mean": 0.0018731327123532537, "completions/clipped_ratio": 0.0, "completions/max_length": 567.1, "completions/max_terminated_length": 567.1, "completions/mean_length": 186.7916717529297, "completions/mean_terminated_length": 186.7916717529297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028858504661669334, "epoch": 15.202020202020202, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 2544011344.0, "reward": 7.9272722721099855, "reward_std": 0.3612960487604141, "rewards/combined_reward/mean": 7.927272176742553, "rewards/combined_reward/std": 2.2660353660583494, "step": 24840 }, { "clip_ratio/high_max": 0.001715154629588748, "clip_ratio/high_mean": 0.0009762076135909107, "clip_ratio/low_mean": 0.0007162301478577623, "clip_ratio/low_min": 0.00017027016486584519, "clip_ratio/region_mean": 0.0016924377540514493, "completions/clipped_ratio": 0.0, "completions/max_length": 596.4, "completions/max_terminated_length": 596.4, "completions/mean_length": 193.61432952880858, "completions/mean_terminated_length": 193.61432952880858, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.026738289541875324, "epoch": 15.22038567493113, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2547054983.0, "reward": 7.998671388626098, "reward_std": 0.38138749450445175, "rewards/combined_reward/mean": 7.998671150207519, "rewards/combined_reward/std": 2.4491347670555115, "step": 24870 }, { "clip_ratio/high_max": 0.0018789736064112124, "clip_ratio/high_mean": 0.001093302092643474, "clip_ratio/low_mean": 0.0007026065125198026, "clip_ratio/low_min": 0.0001849736739435078, "clip_ratio/region_mean": 0.0017959085933398455, "completions/clipped_ratio": 0.0, "completions/max_length": 590.3, "completions/max_terminated_length": 590.3, "completions/mean_length": 200.85938110351563, "completions/mean_terminated_length": 200.85938110351563, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.02798551128556331, "epoch": 15.238751147842057, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.0252685546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2550126443.0, "reward": 8.179546213150024, "reward_std": 0.34039194881916046, "rewards/combined_reward/mean": 8.179545831680297, "rewards/combined_reward/std": 2.10259627699852, "step": 24900 }, { "clip_ratio/high_max": 0.0020197499747155235, "clip_ratio/high_mean": 0.001141098294465337, "clip_ratio/low_mean": 0.0008490359089288783, "clip_ratio/low_min": 0.0002489286659207816, "clip_ratio/region_mean": 0.0019901342069109282, "completions/clipped_ratio": 0.0, "completions/max_length": 500.3, "completions/max_terminated_length": 500.3, "completions/mean_length": 181.46198425292968, "completions/mean_terminated_length": 181.46198425292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03033799856590728, "epoch": 15.257116620752985, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.0069580078125, "learning_rate": 5e-06, "loss": 0.0027, "num_tokens": 2553123417.0, "reward": 7.946627044677735, "reward_std": 0.40091986060142515, "rewards/combined_reward/mean": 7.946626806259156, "rewards/combined_reward/std": 2.4565948605537415, "step": 24930 }, { "clip_ratio/high_max": 0.0018800928757021515, "clip_ratio/high_mean": 0.001158079995123747, "clip_ratio/low_mean": 0.0008243907945143292, "clip_ratio/low_min": 0.0002556425946143766, "clip_ratio/region_mean": 0.001982470780300597, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 192.82005767822267, "completions/mean_terminated_length": 192.82005767822267, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028917265279839437, "epoch": 15.275482093663912, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2556164006.0, "reward": 8.217748212814332, "reward_std": 0.3627097591757774, "rewards/combined_reward/mean": 8.217748212814332, "rewards/combined_reward/std": 2.201750361919403, "step": 24960 }, { "clip_ratio/high_max": 0.002108980436848166, "clip_ratio/high_mean": 0.001272159634148314, "clip_ratio/low_mean": 0.0007832846213811233, "clip_ratio/low_min": 0.00018908413840108552, "clip_ratio/region_mean": 0.0020554442662008418, "completions/clipped_ratio": 0.0, "completions/max_length": 539.7, "completions/max_terminated_length": 539.7, "completions/mean_length": 190.0455780029297, "completions/mean_terminated_length": 190.0455780029297, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.02979522490253051, "epoch": 15.29384756657484, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.0263671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2559193941.0, "reward": 7.764581394195557, "reward_std": 0.3936672702431679, "rewards/combined_reward/mean": 7.764581346511841, "rewards/combined_reward/std": 2.538193643093109, "step": 24990 }, { "clip_ratio/high_max": 0.0016571539740349787, "clip_ratio/high_mean": 0.0009340698698603471, "clip_ratio/low_mean": 0.0008889375467939924, "clip_ratio/low_min": 0.00032789183799953513, "clip_ratio/region_mean": 0.0018230074415138612, "completions/clipped_ratio": 0.0, "completions/max_length": 649.5, "completions/max_terminated_length": 649.5, "completions/mean_length": 198.31875457763672, "completions/mean_terminated_length": 198.31875457763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03182406093304356, "epoch": 15.312213039485767, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2562255645.0, "reward": 7.926143646240234, "reward_std": 0.30705697536468507, "rewards/combined_reward/mean": 7.926143407821655, "rewards/combined_reward/std": 2.4380086302757262, "step": 25020 }, { "clip_ratio/high_max": 0.0018658702162307843, "clip_ratio/high_mean": 0.0010571433448300618, "clip_ratio/low_mean": 0.0007578509112136089, "clip_ratio/low_min": 0.0002568271542259026, "clip_ratio/region_mean": 0.0018149942775683788, "completions/clipped_ratio": 0.0, "completions/max_length": 503.1, "completions/max_terminated_length": 503.1, "completions/mean_length": 193.1432357788086, "completions/mean_terminated_length": 193.1432357788086, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.028868294685768583, "epoch": 15.330578512396695, "frac_reward_zero_std": 0.1625000022351742, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2565297475.0, "reward": 8.055152463912965, "reward_std": 0.44970834255218506, "rewards/combined_reward/mean": 8.055152130126952, "rewards/combined_reward/std": 2.173030507564545, "step": 25050 }, { "clip_ratio/high_max": 0.0016658912420098205, "clip_ratio/high_mean": 0.0009455621674230012, "clip_ratio/low_mean": 0.0007382420789023551, "clip_ratio/low_min": 0.00017368621192872524, "clip_ratio/region_mean": 0.00168380424050459, "completions/clipped_ratio": 0.0, "completions/max_length": 567.1, "completions/max_terminated_length": 567.1, "completions/mean_length": 203.6408905029297, "completions/mean_terminated_length": 203.6408905029297, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.02760328717219333, "epoch": 15.348943985307622, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2568379616.0, "reward": 8.178805351257324, "reward_std": 0.3245466113090515, "rewards/combined_reward/mean": 8.178805065155029, "rewards/combined_reward/std": 2.150813543796539, "step": 25080 }, { "clip_ratio/high_max": 0.00197789649682818, "clip_ratio/high_mean": 0.0011239932686294197, "clip_ratio/low_mean": 0.0008177781442403405, "clip_ratio/low_min": 0.00022977557455305941, "clip_ratio/region_mean": 0.0019417714094743133, "completions/clipped_ratio": 0.0, "completions/max_length": 547.6, "completions/max_terminated_length": 547.6, "completions/mean_length": 192.04141082763672, "completions/mean_terminated_length": 192.04141082763672, "completions/min_length": 48.2, "completions/min_terminated_length": 48.2, "entropy": 0.02821291956740121, "epoch": 15.36730945821855, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2571417215.0, "reward": 8.15639524459839, "reward_std": 0.4026845172047615, "rewards/combined_reward/mean": 8.156394767761231, "rewards/combined_reward/std": 2.08501877784729, "step": 25110 }, { "clip_ratio/high_max": 0.0025150936140562407, "clip_ratio/high_mean": 0.0012551749099657173, "clip_ratio/low_mean": 0.0007560297396897416, "clip_ratio/low_min": 0.00018613027762815667, "clip_ratio/region_mean": 0.002011204646260012, "completions/clipped_ratio": 0.0, "completions/max_length": 594.3, "completions/max_terminated_length": 594.3, "completions/mean_length": 202.48985137939454, "completions/mean_terminated_length": 202.48985137939454, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.030950648663565516, "epoch": 15.385674931129477, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2574494936.0, "reward": 7.888097858428955, "reward_std": 0.38476870208978653, "rewards/combined_reward/mean": 7.888097667694092, "rewards/combined_reward/std": 2.3969976782798765, "step": 25140 }, { "clip_ratio/high_max": 0.001985629044793313, "clip_ratio/high_mean": 0.0010793372628540964, "clip_ratio/low_mean": 0.0008440411240978089, "clip_ratio/low_min": 0.00022204975757631474, "clip_ratio/region_mean": 0.0019233783837989903, "completions/clipped_ratio": 0.0, "completions/max_length": 560.6, "completions/max_terminated_length": 560.6, "completions/mean_length": 203.8424530029297, "completions/mean_terminated_length": 203.8424530029297, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02888815539578597, "epoch": 15.404040404040405, "frac_reward_zero_std": 0.15833333916962147, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2577577851.0, "reward": 8.234542798995971, "reward_std": 0.2887293815612793, "rewards/combined_reward/mean": 8.23454246520996, "rewards/combined_reward/std": 2.1317402720451355, "step": 25170 }, { "clip_ratio/high_max": 0.0018055338812700938, "clip_ratio/high_mean": 0.001099940411586431, "clip_ratio/low_mean": 0.0008706122876295315, "clip_ratio/low_min": 0.00024780076831423987, "clip_ratio/region_mean": 0.0019705526797527757, "completions/clipped_ratio": 0.0, "completions/max_length": 634.4, "completions/max_terminated_length": 634.4, "completions/mean_length": 200.94427490234375, "completions/mean_terminated_length": 200.94427490234375, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.03240147723505894, "epoch": 15.42240587695133, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2580649637.0, "reward": 8.095155668258666, "reward_std": 0.3643873929977417, "rewards/combined_reward/mean": 8.095155620574952, "rewards/combined_reward/std": 2.4254923701286315, "step": 25200 }, { "clip_ratio/high_max": 0.0020379922139303138, "clip_ratio/high_mean": 0.0012083246983820572, "clip_ratio/low_mean": 0.0007975856703221021, "clip_ratio/low_min": 0.00021600301430832285, "clip_ratio/region_mean": 0.0020059103689466914, "completions/clipped_ratio": 0.0, "completions/max_length": 609.8, "completions/max_terminated_length": 609.8, "completions/mean_length": 202.61302642822267, "completions/mean_terminated_length": 202.61302642822267, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03226597138370077, "epoch": 15.44077134986226, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2583727831.0, "reward": 8.061291408538818, "reward_std": 0.3440116584300995, "rewards/combined_reward/mean": 8.061291027069093, "rewards/combined_reward/std": 2.3991721630096436, "step": 25230 }, { "clip_ratio/high_max": 0.0016062582731440973, "clip_ratio/high_mean": 0.0008607831521658226, "clip_ratio/low_mean": 0.0006892915428276562, "clip_ratio/low_min": 0.0002181029611771616, "clip_ratio/region_mean": 0.0015500747046947557, "completions/clipped_ratio": 0.0, "completions/max_length": 586.1, "completions/max_terminated_length": 586.1, "completions/mean_length": 203.8612045288086, "completions/mean_terminated_length": 203.8612045288086, "completions/min_length": 48.4, "completions/min_terminated_length": 48.4, "entropy": 0.0248208539482827, "epoch": 15.459136822773186, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.00762939453125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2586810818.0, "reward": 8.14604573249817, "reward_std": 0.31723438054323194, "rewards/combined_reward/mean": 8.14604549407959, "rewards/combined_reward/std": 2.1176859378814696, "step": 25260 }, { "clip_ratio/high_max": 0.0018213972623925656, "clip_ratio/high_mean": 0.0010240805266221286, "clip_ratio/low_mean": 0.0008651184648139558, "clip_ratio/low_min": 0.00020420610538470405, "clip_ratio/region_mean": 0.00188919898646418, "completions/clipped_ratio": 0.0, "completions/max_length": 649.8, "completions/max_terminated_length": 649.8, "completions/mean_length": 197.13906860351562, "completions/mean_terminated_length": 197.13906860351562, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02921581723106404, "epoch": 15.477502295684113, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.01422119140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2589867992.0, "reward": 7.799942493438721, "reward_std": 0.37960695922374726, "rewards/combined_reward/mean": 7.799942302703857, "rewards/combined_reward/std": 2.509660816192627, "step": 25290 }, { "clip_ratio/high_max": 0.0019371333527184711, "clip_ratio/high_mean": 0.0010626968345604838, "clip_ratio/low_mean": 0.0007040307045826921, "clip_ratio/low_min": 0.00018179115674380833, "clip_ratio/region_mean": 0.001766727535383931, "completions/clipped_ratio": 0.0, "completions/max_length": 552.8, "completions/max_terminated_length": 552.8, "completions/mean_length": 199.30469207763673, "completions/mean_terminated_length": 199.30469207763673, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.028302050249961515, "epoch": 15.49586776859504, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2592933482.0, "reward": 8.15032114982605, "reward_std": 0.3044921740889549, "rewards/combined_reward/mean": 8.150321102142334, "rewards/combined_reward/std": 2.1584079027175904, "step": 25320 }, { "clip_ratio/high_max": 0.0016605432101641782, "clip_ratio/high_mean": 0.0009765470427131125, "clip_ratio/low_mean": 0.0006827856999734649, "clip_ratio/low_min": 0.00013895906522520818, "clip_ratio/region_mean": 0.001659332756389631, "completions/clipped_ratio": 0.0, "completions/max_length": 654.4, "completions/max_terminated_length": 654.4, "completions/mean_length": 196.30677642822266, "completions/mean_terminated_length": 196.30677642822266, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.02595855507533997, "epoch": 15.514233241505968, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2595987460.0, "reward": 8.032768487930298, "reward_std": 0.358998116850853, "rewards/combined_reward/mean": 8.03276834487915, "rewards/combined_reward/std": 2.209522914886475, "step": 25350 }, { "clip_ratio/high_max": 0.0019807500580403333, "clip_ratio/high_mean": 0.0010696377675534071, "clip_ratio/low_mean": 0.0008041208437740958, "clip_ratio/low_min": 0.0002036459790057658, "clip_ratio/region_mean": 0.0018737586059918007, "completions/clipped_ratio": 0.0, "completions/max_length": 526.7, "completions/max_terminated_length": 526.7, "completions/mean_length": 197.07031707763673, "completions/mean_terminated_length": 197.07031707763673, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03016483342895905, "epoch": 15.532598714416896, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2599044370.0, "reward": 8.068010663986206, "reward_std": 0.37991664558649063, "rewards/combined_reward/mean": 8.068010330200195, "rewards/combined_reward/std": 2.3902955174446108, "step": 25380 }, { "clip_ratio/high_max": 0.0020471724565140903, "clip_ratio/high_mean": 0.0011737282130828439, "clip_ratio/low_mean": 0.0006813213796704077, "clip_ratio/low_min": 0.00021620634846234073, "clip_ratio/region_mean": 0.0018550495903279323, "completions/clipped_ratio": 0.0, "completions/max_length": 587.3, "completions/max_terminated_length": 587.3, "completions/mean_length": 191.6432342529297, "completions/mean_terminated_length": 191.6432342529297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02708151946620395, "epoch": 15.550964187327823, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.0361328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2602080440.0, "reward": 7.941785621643066, "reward_std": 0.3520200863480568, "rewards/combined_reward/mean": 7.941785669326782, "rewards/combined_reward/std": 2.303845930099487, "step": 25410 }, { "clip_ratio/high_max": 0.0024116925424702155, "clip_ratio/high_mean": 0.0014439441251549094, "clip_ratio/low_mean": 0.0007099503173473447, "clip_ratio/low_min": 0.0001841315177443903, "clip_ratio/region_mean": 0.0021538944740314035, "completions/clipped_ratio": 0.0, "completions/max_length": 670.4, "completions/max_terminated_length": 670.4, "completions/mean_length": 211.20677337646484, "completions/mean_terminated_length": 211.20677337646484, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.033155119760582843, "epoch": 15.56932966023875, "frac_reward_zero_std": 0.1291666690260172, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2605191634.0, "reward": 7.9953944206237795, "reward_std": 0.3593493938446045, "rewards/combined_reward/mean": 7.995394134521485, "rewards/combined_reward/std": 2.2558564066886904, "step": 25440 }, { "clip_ratio/high_max": 0.0019963303318945693, "clip_ratio/high_mean": 0.0012176257516936554, "clip_ratio/low_mean": 0.000896671390364645, "clip_ratio/low_min": 0.00031200971158493, "clip_ratio/region_mean": 0.002114297134297279, "completions/clipped_ratio": 0.0, "completions/max_length": 635.6, "completions/max_terminated_length": 635.6, "completions/mean_length": 209.97240142822267, "completions/mean_terminated_length": 209.97240142822267, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.03163816697585086, "epoch": 15.587695133149678, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2608298088.0, "reward": 8.160098266601562, "reward_std": 0.3364489316940308, "rewards/combined_reward/mean": 8.160097980499268, "rewards/combined_reward/std": 2.2246580600738524, "step": 25470 }, { "clip_ratio/high_max": 0.0021674402790570944, "clip_ratio/high_mean": 0.0013155741603744294, "clip_ratio/low_mean": 0.0007234190101977826, "clip_ratio/low_min": 0.00021755064323466893, "clip_ratio/region_mean": 0.002038993149471935, "completions/clipped_ratio": 0.0, "completions/max_length": 589.9, "completions/max_terminated_length": 589.9, "completions/mean_length": 202.11693115234374, "completions/mean_terminated_length": 202.11693115234374, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02966462665547927, "epoch": 15.606060606060606, "frac_reward_zero_std": 0.18750000633299352, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2611374377.0, "reward": 8.111000633239746, "reward_std": 0.3773253381252289, "rewards/combined_reward/mean": 8.111000490188598, "rewards/combined_reward/std": 2.1827357172966004, "step": 25500 }, { "clip_ratio/high_max": 0.002003898443945218, "clip_ratio/high_mean": 0.0011439898302342045, "clip_ratio/low_mean": 0.000821483143226942, "clip_ratio/low_min": 0.0001473254495067522, "clip_ratio/region_mean": 0.0019654729544223907, "completions/clipped_ratio": 0.0, "completions/max_length": 583.1, "completions/max_terminated_length": 583.1, "completions/mean_length": 193.16380462646484, "completions/mean_terminated_length": 193.16380462646484, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02851471131046613, "epoch": 15.624426078971533, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2614416286.0, "reward": 7.810232496261596, "reward_std": 0.3254135325551033, "rewards/combined_reward/mean": 7.810232305526734, "rewards/combined_reward/std": 2.4707339286804197, "step": 25530 }, { "clip_ratio/high_max": 0.0018376989886746743, "clip_ratio/high_mean": 0.0009735647230021035, "clip_ratio/low_mean": 0.0009128995430122207, "clip_ratio/low_min": 0.00034701428691429704, "clip_ratio/region_mean": 0.0018864642746242073, "completions/clipped_ratio": 0.0, "completions/max_length": 519.8, "completions/max_terminated_length": 519.8, "completions/mean_length": 193.9830795288086, "completions/mean_terminated_length": 193.9830795288086, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.028118151426315307, "epoch": 15.64279155188246, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2617461341.0, "reward": 8.050703382492065, "reward_std": 0.3383431568741798, "rewards/combined_reward/mean": 8.050703239440917, "rewards/combined_reward/std": 2.172149419784546, "step": 25560 }, { "clip_ratio/high_max": 0.0019488701786031016, "clip_ratio/high_mean": 0.001112172931122283, "clip_ratio/low_mean": 0.000791302008292405, "clip_ratio/low_min": 0.0002156999895911819, "clip_ratio/region_mean": 0.0019034749483883692, "completions/clipped_ratio": 0.0, "completions/max_length": 637.4, "completions/max_terminated_length": 637.4, "completions/mean_length": 196.79271392822267, "completions/mean_terminated_length": 196.79271392822267, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028041953065743048, "epoch": 15.661157024793388, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2620517185.0, "reward": 8.179954862594604, "reward_std": 0.3096440494060516, "rewards/combined_reward/mean": 8.179954719543456, "rewards/combined_reward/std": 2.2922232866287233, "step": 25590 }, { "clip_ratio/high_max": 0.0016241487091368375, "clip_ratio/high_mean": 0.0008575222913350444, "clip_ratio/low_mean": 0.0007615632956003537, "clip_ratio/low_min": 0.00018539914866172086, "clip_ratio/region_mean": 0.0016190855740812063, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 183.4362030029297, "completions/mean_terminated_length": 183.4362030029297, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.026523846255925795, "epoch": 15.679522497704315, "frac_reward_zero_std": 0.23750000447034836, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2623521740.0, "reward": 7.967030429840088, "reward_std": 0.2830955907702446, "rewards/combined_reward/mean": 7.96703028678894, "rewards/combined_reward/std": 2.2518588066101075, "step": 25620 }, { "clip_ratio/high_max": 0.0021750468949903734, "clip_ratio/high_mean": 0.0011780288734977754, "clip_ratio/low_mean": 0.000642384159436915, "clip_ratio/low_min": 0.00015889157657511532, "clip_ratio/region_mean": 0.001820413037300265, "completions/clipped_ratio": 0.0, "completions/max_length": 545.8, "completions/max_terminated_length": 545.8, "completions/mean_length": 189.69375610351562, "completions/mean_terminated_length": 189.69375610351562, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.030059647963692745, "epoch": 15.697887970615243, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0223388671875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2626550324.0, "reward": 8.10011076927185, "reward_std": 0.2853624030947685, "rewards/combined_reward/mean": 8.10011076927185, "rewards/combined_reward/std": 2.368494713306427, "step": 25650 }, { "clip_ratio/high_max": 0.0020455992130640273, "clip_ratio/high_mean": 0.001249244436742932, "clip_ratio/low_mean": 0.0007868616402750679, "clip_ratio/low_min": 0.0002330764036742039, "clip_ratio/region_mean": 0.002036106095571692, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 203.2330780029297, "completions/mean_terminated_length": 203.2330780029297, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.031413153558969495, "epoch": 15.71625344352617, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 2629630899.0, "reward": 7.983034086227417, "reward_std": 0.35882840007543565, "rewards/combined_reward/mean": 7.983033895492554, "rewards/combined_reward/std": 2.3806650757789614, "step": 25680 }, { "clip_ratio/high_max": 0.0014192375369020739, "clip_ratio/high_mean": 0.0008286272054344106, "clip_ratio/low_mean": 0.0006857321939605753, "clip_ratio/low_min": 0.00015732518004369923, "clip_ratio/region_mean": 0.0015143594129767735, "completions/clipped_ratio": 0.0, "completions/max_length": 553.2, "completions/max_terminated_length": 553.2, "completions/mean_length": 194.6721405029297, "completions/mean_terminated_length": 194.6721405029297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.025396301170500617, "epoch": 15.734618916437098, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2632678600.0, "reward": 7.9339391708374025, "reward_std": 0.28837121576070784, "rewards/combined_reward/mean": 7.933938884735108, "rewards/combined_reward/std": 2.3839831709861756, "step": 25710 }, { "clip_ratio/high_max": 0.0018596468619458998, "clip_ratio/high_mean": 0.001061735546560764, "clip_ratio/low_mean": 0.0007318983595420528, "clip_ratio/low_min": 0.00023925266456596242, "clip_ratio/region_mean": 0.0017936339233225833, "completions/clipped_ratio": 0.0, "completions/max_length": 598.9, "completions/max_terminated_length": 598.9, "completions/mean_length": 206.16250762939453, "completions/mean_terminated_length": 206.16250762939453, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02810027082450688, "epoch": 15.752984389348025, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.0341796875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2635770424.0, "reward": 8.28695559501648, "reward_std": 0.28543743640184405, "rewards/combined_reward/mean": 8.286955261230469, "rewards/combined_reward/std": 1.98407062292099, "step": 25740 }, { "clip_ratio/high_max": 0.0018225658781981717, "clip_ratio/high_mean": 0.001033592845245342, "clip_ratio/low_mean": 0.000876843131967083, "clip_ratio/low_min": 0.0001926714952181404, "clip_ratio/region_mean": 0.0019104359855797763, "completions/clipped_ratio": 0.0, "completions/max_length": 535.1, "completions/max_terminated_length": 535.1, "completions/mean_length": 183.79063262939454, "completions/mean_terminated_length": 183.79063262939454, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028315132406229775, "epoch": 15.771349862258953, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2638776340.0, "reward": 8.249747514724731, "reward_std": 0.2971131667494774, "rewards/combined_reward/mean": 8.249747133255005, "rewards/combined_reward/std": 2.08546906709671, "step": 25770 }, { "clip_ratio/high_max": 0.0018338397863165786, "clip_ratio/high_mean": 0.0010078895039138539, "clip_ratio/low_mean": 0.0006981488229939714, "clip_ratio/low_min": 0.000178166540475407, "clip_ratio/region_mean": 0.0017060383236336444, "completions/clipped_ratio": 0.0, "completions/max_length": 541.5, "completions/max_terminated_length": 541.5, "completions/mean_length": 198.9705780029297, "completions/mean_terminated_length": 198.9705780029297, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.027559348521754144, "epoch": 15.78971533516988, "frac_reward_zero_std": 0.14583333879709243, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2641840547.0, "reward": 8.049615669250489, "reward_std": 0.3592381328344345, "rewards/combined_reward/mean": 8.049615240097046, "rewards/combined_reward/std": 2.3186096429824827, "step": 25800 }, { "clip_ratio/high_max": 0.0021136968663389172, "clip_ratio/high_mean": 0.0011688464096854053, "clip_ratio/low_mean": 0.0007374108407627015, "clip_ratio/low_min": 0.0002525316464016214, "clip_ratio/region_mean": 0.0019062572629385008, "completions/clipped_ratio": 0.0, "completions/max_length": 561.1, "completions/max_terminated_length": 561.1, "completions/mean_length": 206.4875030517578, "completions/mean_terminated_length": 206.4875030517578, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.029725847548494735, "epoch": 15.808080808080808, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.0068359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2644933619.0, "reward": 8.34679627418518, "reward_std": 0.3062371641397476, "rewards/combined_reward/mean": 8.346795988082885, "rewards/combined_reward/std": 1.964946174621582, "step": 25830 }, { "clip_ratio/high_max": 0.0021922433360790214, "clip_ratio/high_mean": 0.0011215859893127344, "clip_ratio/low_mean": 0.0006745805648582366, "clip_ratio/low_min": 0.00016653434649924748, "clip_ratio/region_mean": 0.0017961665527157797, "completions/clipped_ratio": 0.0, "completions/max_length": 509.6, "completions/max_terminated_length": 509.6, "completions/mean_length": 195.0487045288086, "completions/mean_terminated_length": 195.0487045288086, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027816563363497457, "epoch": 15.826446280991735, "frac_reward_zero_std": 0.20833334177732468, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2647982766.0, "reward": 8.006843996047973, "reward_std": 0.338391549885273, "rewards/combined_reward/mean": 8.00684380531311, "rewards/combined_reward/std": 2.2623153567314147, "step": 25860 }, { "clip_ratio/high_max": 0.0022029549402456422, "clip_ratio/high_mean": 0.0012121795751833511, "clip_ratio/low_mean": 0.0007717365384451114, "clip_ratio/low_min": 0.0001879344975653415, "clip_ratio/region_mean": 0.001983916142489761, "completions/clipped_ratio": 0.0, "completions/max_length": 647.3, "completions/max_terminated_length": 647.3, "completions/mean_length": 189.04896392822266, "completions/mean_terminated_length": 189.04896392822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028364694250437122, "epoch": 15.844811753902663, "frac_reward_zero_std": 0.23333334028720856, "grad_norm": 0.006103515625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2651008874.0, "reward": 8.100493144989013, "reward_std": 0.3452778160572052, "rewards/combined_reward/mean": 8.100493001937867, "rewards/combined_reward/std": 2.1751306295394897, "step": 25890 }, { "clip_ratio/high_max": 0.0019528017612174154, "clip_ratio/high_mean": 0.0011625863374016868, "clip_ratio/low_mean": 0.0006656137925650303, "clip_ratio/low_min": 0.0001598575385287404, "clip_ratio/region_mean": 0.0018282001134745467, "completions/clipped_ratio": 0.0, "completions/max_length": 611.6, "completions/max_terminated_length": 611.6, "completions/mean_length": 203.57969360351564, "completions/mean_terminated_length": 203.57969360351564, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.031153335437799492, "epoch": 15.86317722681359, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.03076171875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2654090780.0, "reward": 8.27339277267456, "reward_std": 0.36625549495220183, "rewards/combined_reward/mean": 8.273392629623412, "rewards/combined_reward/std": 2.0477567672729493, "step": 25920 }, { "clip_ratio/high_max": 0.0019987332042849934, "clip_ratio/high_mean": 0.0011500064429128543, "clip_ratio/low_mean": 0.0007755110467163225, "clip_ratio/low_min": 0.0002905346812137092, "clip_ratio/region_mean": 0.001925517501270709, "completions/clipped_ratio": 0.0, "completions/max_length": 589.3, "completions/max_terminated_length": 589.3, "completions/mean_length": 191.53880920410157, "completions/mean_terminated_length": 191.53880920410157, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03071690248325467, "epoch": 15.881542699724518, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2657126449.0, "reward": 8.188122224807739, "reward_std": 0.3221184104681015, "rewards/combined_reward/mean": 8.188122034072876, "rewards/combined_reward/std": 2.1367731928825378, "step": 25950 }, { "clip_ratio/high_max": 0.0021346626075683162, "clip_ratio/high_mean": 0.001144428575450244, "clip_ratio/low_mean": 0.0006830427758056127, "clip_ratio/low_min": 0.00019309182947229903, "clip_ratio/region_mean": 0.0018274713521047184, "completions/clipped_ratio": 0.0, "completions/max_length": 586.1, "completions/max_terminated_length": 586.1, "completions/mean_length": 203.0132843017578, "completions/mean_terminated_length": 203.0132843017578, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.029710323619656264, "epoch": 15.899908172635445, "frac_reward_zero_std": 0.14166667126119137, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2660206180.0, "reward": 8.059128379821777, "reward_std": 0.3740886151790619, "rewards/combined_reward/mean": 8.059128141403198, "rewards/combined_reward/std": 2.2390969514846804, "step": 25980 }, { "clip_ratio/high_max": 0.0019460350842564367, "clip_ratio/high_mean": 0.0011356212035025237, "clip_ratio/low_mean": 0.000777403543421921, "clip_ratio/low_min": 0.00022057343982548143, "clip_ratio/region_mean": 0.0019130247499560937, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 196.1979232788086, "completions/mean_terminated_length": 196.1979232788086, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.030976494153340656, "epoch": 15.918273645546373, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.005615234375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2663259740.0, "reward": 8.061134243011475, "reward_std": 0.29690295457839966, "rewards/combined_reward/mean": 8.06113405227661, "rewards/combined_reward/std": 2.382638621330261, "step": 26010 }, { "clip_ratio/high_max": 0.001754870459990343, "clip_ratio/high_mean": 0.001057948302332079, "clip_ratio/low_mean": 0.0007337779443332693, "clip_ratio/low_min": 0.0002116051357006654, "clip_ratio/region_mean": 0.001791726217682784, "completions/clipped_ratio": 0.0, "completions/max_length": 512.6, "completions/max_terminated_length": 512.6, "completions/mean_length": 180.59089050292968, "completions/mean_terminated_length": 180.59089050292968, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02681819524926444, "epoch": 15.9366391184573, "frac_reward_zero_std": 0.2125000052154064, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2666253369.0, "reward": 7.746448135375976, "reward_std": 0.31958580762147903, "rewards/combined_reward/mean": 7.746448087692261, "rewards/combined_reward/std": 2.543926727771759, "step": 26040 }, { "clip_ratio/high_max": 0.0021341918662074024, "clip_ratio/high_mean": 0.0012235950317214398, "clip_ratio/low_mean": 0.0007939403596537887, "clip_ratio/low_min": 0.0002032039476034697, "clip_ratio/region_mean": 0.0020175353982873883, "completions/clipped_ratio": 0.0, "completions/max_length": 563.1, "completions/max_terminated_length": 563.1, "completions/mean_length": 203.94818267822265, "completions/mean_terminated_length": 203.94818267822265, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.029557170377423368, "epoch": 15.955004591368228, "frac_reward_zero_std": 0.12500000298023223, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2669336690.0, "reward": 8.109821367263795, "reward_std": 0.36287135183811187, "rewards/combined_reward/mean": 8.109820890426636, "rewards/combined_reward/std": 2.074566161632538, "step": 26070 }, { "clip_ratio/high_max": 0.0018426608886026466, "clip_ratio/high_mean": 0.0010712920067210993, "clip_ratio/low_mean": 0.0007194470006652409, "clip_ratio/low_min": 0.00020421760418685154, "clip_ratio/region_mean": 0.0017907389978063294, "completions/clipped_ratio": 0.0, "completions/max_length": 533.8, "completions/max_terminated_length": 533.8, "completions/mean_length": 195.3664108276367, "completions/mean_terminated_length": 195.3664108276367, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02881428039011856, "epoch": 15.973370064279155, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 2672387057.0, "reward": 7.817143964767456, "reward_std": 0.29724001586437226, "rewards/combined_reward/mean": 7.817143774032592, "rewards/combined_reward/std": 2.5451475381851196, "step": 26100 }, { "clip_ratio/high_max": 0.0017130265492596663, "clip_ratio/high_mean": 0.0009340359725077481, "clip_ratio/low_mean": 0.0008461770068000381, "clip_ratio/low_min": 0.0002886296936291425, "clip_ratio/region_mean": 0.001780212989736659, "completions/clipped_ratio": 0.0, "completions/max_length": 644.5, "completions/max_terminated_length": 644.5, "completions/mean_length": 201.3179733276367, "completions/mean_terminated_length": 201.3179733276367, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02773648235015571, "epoch": 15.991735537190083, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2675460278.0, "reward": 7.961833095550537, "reward_std": 0.29523391649127007, "rewards/combined_reward/mean": 7.961832904815674, "rewards/combined_reward/std": 2.3698278427124024, "step": 26130 }, { "clip_ratio/high_max": 0.0022878345487578238, "clip_ratio/high_mean": 0.0013336357370746078, "clip_ratio/low_mean": 0.0008775720845852581, "clip_ratio/low_min": 0.0003154008872578012, "clip_ratio/region_mean": 0.0022112078479010243, "completions/clipped_ratio": 0.0, "completions/max_length": 627.2, "completions/max_terminated_length": 627.2, "completions/mean_length": 203.29192962646485, "completions/mean_terminated_length": 203.29192962646485, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02864582267147107, "epoch": 16.010407101316193, "frac_reward_zero_std": 0.1375000048428774, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2678541079.0, "reward": 8.07959942817688, "reward_std": 0.3038766890764236, "rewards/combined_reward/mean": 8.079599332809448, "rewards/combined_reward/std": 2.11454439163208, "step": 26160 }, { "clip_ratio/high_max": 0.002272756558037751, "clip_ratio/high_mean": 0.0012597441679342107, "clip_ratio/low_mean": 0.0010451820538340446, "clip_ratio/low_min": 0.0003501200136573364, "clip_ratio/region_mean": 0.0023049262087927974, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 199.49557800292968, "completions/mean_terminated_length": 199.49557800292968, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.026943242254977424, "epoch": 16.02877257422712, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2681607302.0, "reward": 8.064150285720824, "reward_std": 0.267789950966835, "rewards/combined_reward/mean": 8.064150047302245, "rewards/combined_reward/std": 2.2442546844482423, "step": 26190 }, { "clip_ratio/high_max": 0.002576021461572964, "clip_ratio/high_mean": 0.0014063494909350993, "clip_ratio/low_mean": 0.0009323928562177268, "clip_ratio/low_min": 0.00023556141774558152, "clip_ratio/region_mean": 0.0023387423396343367, "completions/clipped_ratio": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 205.3351623535156, "completions/mean_terminated_length": 205.3351623535156, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.030268170110260447, "epoch": 16.04713804713805, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0291748046875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2684695949.0, "reward": 7.974419450759887, "reward_std": 0.31351396441459656, "rewards/combined_reward/mean": 7.974419212341308, "rewards/combined_reward/std": 2.4503974676132203, "step": 26220 }, { "clip_ratio/high_max": 0.002778127594986775, "clip_ratio/high_mean": 0.0016785289779363665, "clip_ratio/low_mean": 0.0008629190198917059, "clip_ratio/low_min": 0.00016598940031447757, "clip_ratio/region_mean": 0.002541447966359556, "completions/clipped_ratio": 0.0, "completions/max_length": 624.3, "completions/max_terminated_length": 624.3, "completions/mean_length": 202.7867233276367, "completions/mean_terminated_length": 202.7867233276367, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.031637114689995846, "epoch": 16.065503520048974, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2687774810.0, "reward": 8.088717794418335, "reward_std": 0.31023015528917314, "rewards/combined_reward/mean": 8.08871750831604, "rewards/combined_reward/std": 2.2333565592765807, "step": 26250 }, { "clip_ratio/high_max": 0.0023443895639502443, "clip_ratio/high_mean": 0.0013472304732810396, "clip_ratio/low_mean": 0.001009942832873397, "clip_ratio/low_min": 0.0003529304696712643, "clip_ratio/region_mean": 0.002357173301667596, "completions/clipped_ratio": 0.0, "completions/max_length": 551.5, "completions/max_terminated_length": 551.5, "completions/mean_length": 204.5929748535156, "completions/mean_terminated_length": 204.5929748535156, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.03013729106945296, "epoch": 16.083868992959903, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0036468505859375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2690860607.0, "reward": 7.998499965667724, "reward_std": 0.3046166017651558, "rewards/combined_reward/mean": 7.99849944114685, "rewards/combined_reward/std": 2.338605213165283, "step": 26280 }, { "clip_ratio/high_max": 0.0021175454171801296, "clip_ratio/high_mean": 0.0011788005569542293, "clip_ratio/low_mean": 0.0009630725437091315, "clip_ratio/low_min": 0.0002727260563309149, "clip_ratio/region_mean": 0.0021418730778047274, "completions/clipped_ratio": 0.0, "completions/max_length": 518.7, "completions/max_terminated_length": 518.7, "completions/mean_length": 193.77526397705077, "completions/mean_terminated_length": 193.77526397705077, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.025860008892292778, "epoch": 16.10223446587083, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.0120849609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2693904864.0, "reward": 8.116601705551147, "reward_std": 0.36925110667943956, "rewards/combined_reward/mean": 8.116601467132568, "rewards/combined_reward/std": 2.282694733142853, "step": 26310 }, { "clip_ratio/high_max": 0.002277881122912125, "clip_ratio/high_mean": 0.0013234984335819414, "clip_ratio/low_mean": 0.0009474124543582245, "clip_ratio/low_min": 0.0002686055566300638, "clip_ratio/region_mean": 0.002270910863686974, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 199.0682342529297, "completions/mean_terminated_length": 199.0682342529297, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.026251336823528013, "epoch": 16.12059993878176, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2696969446.0, "reward": 8.288934183120727, "reward_std": 0.2720553562045097, "rewards/combined_reward/mean": 8.288933801651002, "rewards/combined_reward/std": 2.0692915320396423, "step": 26340 }, { "clip_ratio/high_max": 0.002140657472530923, "clip_ratio/high_mean": 0.0012214580537450578, "clip_ratio/low_mean": 0.000982564361159651, "clip_ratio/low_min": 0.0003043017112455952, "clip_ratio/region_mean": 0.0022040224124793895, "completions/clipped_ratio": 0.0, "completions/max_length": 613.4, "completions/max_terminated_length": 613.4, "completions/mean_length": 206.40052490234376, "completions/mean_terminated_length": 206.40052490234376, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.027997823253584405, "epoch": 16.138965411692684, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0050048828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2700062184.0, "reward": 8.143565607070922, "reward_std": 0.3253671944141388, "rewards/combined_reward/mean": 8.14356551170349, "rewards/combined_reward/std": 2.0913867950439453, "step": 26370 }, { "clip_ratio/high_max": 0.003193481250006395, "clip_ratio/high_mean": 0.0019182801139929022, "clip_ratio/low_mean": 0.0011306584416161058, "clip_ratio/low_min": 0.0003051116390755245, "clip_ratio/region_mean": 0.0030489385598533167, "completions/clipped_ratio": 0.0, "completions/max_length": 705.4, "completions/max_terminated_length": 705.4, "completions/mean_length": 194.19531707763673, "completions/mean_terminated_length": 194.19531707763673, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03940144379933675, "epoch": 16.157330884603613, "frac_reward_zero_std": 0.1708333406597376, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2703108054.0, "reward": 8.108645153045654, "reward_std": 0.3640234336256981, "rewards/combined_reward/mean": 8.108645343780518, "rewards/combined_reward/std": 2.173184025287628, "step": 26400 }, { "clip_ratio/high_max": 0.0026901967294785814, "clip_ratio/high_mean": 0.0015267209620409025, "clip_ratio/low_mean": 0.001011462505266536, "clip_ratio/low_min": 0.00018074029164078336, "clip_ratio/region_mean": 0.0025381834789489706, "completions/clipped_ratio": 0.0, "completions/max_length": 593.7, "completions/max_terminated_length": 593.7, "completions/mean_length": 190.8299530029297, "completions/mean_terminated_length": 190.8299530029297, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03168504652567208, "epoch": 16.17569635751454, "frac_reward_zero_std": 0.21666667312383653, "grad_norm": 0.005340576171875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2706141001.0, "reward": 7.582441139221191, "reward_std": 0.33002363592386247, "rewards/combined_reward/mean": 7.582440948486328, "rewards/combined_reward/std": 2.7224831700325014, "step": 26430 }, { "clip_ratio/high_max": 0.0021348910675442314, "clip_ratio/high_mean": 0.0011335801796424978, "clip_ratio/low_mean": 0.0008463318898672393, "clip_ratio/low_min": 0.0002248440692104244, "clip_ratio/region_mean": 0.0019799120656292264, "completions/clipped_ratio": 0.0, "completions/max_length": 546.7, "completions/max_terminated_length": 546.7, "completions/mean_length": 197.16588745117187, "completions/mean_terminated_length": 197.16588745117187, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.025563701700108746, "epoch": 16.19406183042547, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2709198278.0, "reward": 7.970092058181763, "reward_std": 0.3280294701457024, "rewards/combined_reward/mean": 7.970091676712036, "rewards/combined_reward/std": 2.6056381106376647, "step": 26460 }, { "clip_ratio/high_max": 0.002336786635472284, "clip_ratio/high_mean": 0.0012794675236364127, "clip_ratio/low_mean": 0.0009319399757562981, "clip_ratio/low_min": 0.0002258860023478822, "clip_ratio/region_mean": 0.002211407498786381, "completions/clipped_ratio": 0.0, "completions/max_length": 555.3, "completions/max_terminated_length": 555.3, "completions/mean_length": 197.23281860351562, "completions/mean_terminated_length": 197.23281860351562, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02727342173457146, "epoch": 16.212427303336394, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.005859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2712255812.0, "reward": 8.250736093521118, "reward_std": 0.3336705893278122, "rewards/combined_reward/mean": 8.250735759735107, "rewards/combined_reward/std": 2.0718210697174073, "step": 26490 }, { "clip_ratio/high_max": 0.002614066321984865, "clip_ratio/high_mean": 0.0015097798464315322, "clip_ratio/low_mean": 0.0009645317547741191, "clip_ratio/low_min": 0.0002703250749618746, "clip_ratio/region_mean": 0.0024743116065413535, "completions/clipped_ratio": 0.0, "completions/max_length": 532.2, "completions/max_terminated_length": 532.2, "completions/mean_length": 198.13229675292968, "completions/mean_terminated_length": 198.13229675292968, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.029483660721840957, "epoch": 16.230792776247323, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2715316800.0, "reward": 7.87737512588501, "reward_std": 0.3020879432559013, "rewards/combined_reward/mean": 7.877374935150146, "rewards/combined_reward/std": 2.4221477150917052, "step": 26520 }, { "clip_ratio/high_max": 0.0020695570257278934, "clip_ratio/high_mean": 0.0011970544676538945, "clip_ratio/low_mean": 0.000966431627481749, "clip_ratio/low_min": 0.00031598705196908365, "clip_ratio/region_mean": 0.002163486102654133, "completions/clipped_ratio": 0.0, "completions/max_length": 586.4, "completions/max_terminated_length": 586.4, "completions/mean_length": 196.7502670288086, "completions/mean_terminated_length": 196.7502670288086, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.028183597155536214, "epoch": 16.24915824915825, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2718372481.0, "reward": 8.131066608428956, "reward_std": 0.31199254393577575, "rewards/combined_reward/mean": 8.13106632232666, "rewards/combined_reward/std": 2.235110104084015, "step": 26550 }, { "clip_ratio/high_max": 0.002304380466133201, "clip_ratio/high_mean": 0.0013269113892116972, "clip_ratio/low_mean": 0.0009732002467596128, "clip_ratio/low_min": 0.0002824167534223913, "clip_ratio/region_mean": 0.0023001116322120653, "completions/clipped_ratio": 0.0, "completions/max_length": 553.4, "completions/max_terminated_length": 553.4, "completions/mean_length": 199.59219360351562, "completions/mean_terminated_length": 199.59219360351562, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.0282435678721716, "epoch": 16.26752372206918, "frac_reward_zero_std": 0.17500000223517417, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2721439075.0, "reward": 8.363690519332886, "reward_std": 0.318779219686985, "rewards/combined_reward/mean": 8.363690280914307, "rewards/combined_reward/std": 1.8292175769805907, "step": 26580 }, { "clip_ratio/high_max": 0.002089405550941592, "clip_ratio/high_mean": 0.0011371992348964946, "clip_ratio/low_mean": 0.0008765761130537915, "clip_ratio/low_min": 0.000243982042108352, "clip_ratio/region_mean": 0.002013775358500425, "completions/clipped_ratio": 0.0, "completions/max_length": 548.5, "completions/max_terminated_length": 548.5, "completions/mean_length": 197.3807342529297, "completions/mean_terminated_length": 197.3807342529297, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.02720223517778019, "epoch": 16.285889194980104, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2724497177.0, "reward": 8.167417907714844, "reward_std": 0.3532793536782265, "rewards/combined_reward/mean": 8.167417526245117, "rewards/combined_reward/std": 2.111214292049408, "step": 26610 }, { "clip_ratio/high_max": 0.002105198410572484, "clip_ratio/high_mean": 0.001175623924670314, "clip_ratio/low_mean": 0.0009156080736526444, "clip_ratio/low_min": 0.0002454143012679803, "clip_ratio/region_mean": 0.00209123199407865, "completions/clipped_ratio": 0.0, "completions/max_length": 609.4, "completions/max_terminated_length": 609.4, "completions/mean_length": 187.20417175292968, "completions/mean_terminated_length": 187.20417175292968, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02723849721563359, "epoch": 16.30425466789103, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2727516201.0, "reward": 8.27280921936035, "reward_std": 0.2752876952290535, "rewards/combined_reward/mean": 8.272808980941772, "rewards/combined_reward/std": 1.9876643538475036, "step": 26640 }, { "clip_ratio/high_max": 0.0027313038168358616, "clip_ratio/high_mean": 0.0014809838923004767, "clip_ratio/low_mean": 0.0008720202917174902, "clip_ratio/low_min": 0.00021501937444554642, "clip_ratio/region_mean": 0.002353004186685818, "completions/clipped_ratio": 0.0, "completions/max_length": 611.8, "completions/max_terminated_length": 611.8, "completions/mean_length": 202.0187545776367, "completions/mean_terminated_length": 202.0187545776367, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.02984660203413417, "epoch": 16.32262014080196, "frac_reward_zero_std": 0.14583333544433116, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2730592113.0, "reward": 7.949334049224854, "reward_std": 0.3561135396361351, "rewards/combined_reward/mean": 7.949333667755127, "rewards/combined_reward/std": 2.4012121200561523, "step": 26670 }, { "clip_ratio/high_max": 0.00229118674712178, "clip_ratio/high_mean": 0.0012110801211141127, "clip_ratio/low_mean": 0.0010003785326868333, "clip_ratio/low_min": 0.00027078842467744837, "clip_ratio/region_mean": 0.0022114586555593025, "completions/clipped_ratio": 0.0, "completions/max_length": 527.2, "completions/max_terminated_length": 527.2, "completions/mean_length": 204.3315170288086, "completions/mean_terminated_length": 204.3315170288086, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.02759030197436611, "epoch": 16.340985613712885, "frac_reward_zero_std": 0.13333333507180214, "grad_norm": 0.0247802734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2733676906.0, "reward": 8.010520792007446, "reward_std": 0.3243452735245228, "rewards/combined_reward/mean": 8.010520553588867, "rewards/combined_reward/std": 2.524375486373901, "step": 26700 }, { "clip_ratio/high_max": 0.002933105411405753, "clip_ratio/high_mean": 0.001558639356517233, "clip_ratio/low_mean": 0.0009213373550058652, "clip_ratio/low_min": 0.00026049719538908295, "clip_ratio/region_mean": 0.0024799767023068854, "completions/clipped_ratio": 0.0, "completions/max_length": 667.2, "completions/max_terminated_length": 667.2, "completions/mean_length": 207.2375045776367, "completions/mean_terminated_length": 207.2375045776367, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.0308324850940456, "epoch": 16.359351086623814, "frac_reward_zero_std": 0.15833333507180214, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2736772858.0, "reward": 8.070708084106446, "reward_std": 0.33081581741571425, "rewards/combined_reward/mean": 8.070707416534423, "rewards/combined_reward/std": 2.202944040298462, "step": 26730 }, { "clip_ratio/high_max": 0.002586231751774903, "clip_ratio/high_mean": 0.0015301586502270462, "clip_ratio/low_mean": 0.001032845715599251, "clip_ratio/low_min": 0.0003190888931082251, "clip_ratio/region_mean": 0.002563004384379989, "completions/clipped_ratio": 0.0, "completions/max_length": 566.5, "completions/max_terminated_length": 566.5, "completions/mean_length": 197.84479522705078, "completions/mean_terminated_length": 197.84479522705078, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.028510240418836475, "epoch": 16.37771655953474, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2739832742.0, "reward": 8.32193627357483, "reward_std": 0.2795877784490585, "rewards/combined_reward/mean": 8.321935510635376, "rewards/combined_reward/std": 1.8565439403057098, "step": 26760 }, { "clip_ratio/high_max": 0.00258394339713656, "clip_ratio/high_mean": 0.0015060412690218073, "clip_ratio/low_mean": 0.0008886797782300467, "clip_ratio/low_min": 0.00022637895550966884, "clip_ratio/region_mean": 0.002394721071080615, "completions/clipped_ratio": 0.0, "completions/max_length": 572.8, "completions/max_terminated_length": 572.8, "completions/mean_length": 192.01172485351563, "completions/mean_terminated_length": 192.01172485351563, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026558756781741976, "epoch": 16.39608203244567, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.025146484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2742870227.0, "reward": 7.987082290649414, "reward_std": 0.3343091785907745, "rewards/combined_reward/mean": 7.987082195281983, "rewards/combined_reward/std": 2.3117360353469847, "step": 26790 }, { "clip_ratio/high_max": 0.002092184927460039, "clip_ratio/high_mean": 0.001159941222431371, "clip_ratio/low_mean": 0.0008493754772644025, "clip_ratio/low_min": 0.00028253326745470987, "clip_ratio/region_mean": 0.0020093166902370287, "completions/clipped_ratio": 0.0, "completions/max_length": 682.1428571428571, "completions/max_terminated_length": 682.1428571428571, "completions/mean_length": 210.77865164620536, "completions/mean_terminated_length": 210.77865164620536, "completions/min_length": 42.42857142857143, "completions/min_terminated_length": 42.42857142857143, "entropy": 0.034631283255293965, "epoch": 16.413835322926232, "frac_reward_zero_std": 0.16666667269808905, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 2745974000.0, "reward": 8.258497783115931, "reward_std": 0.3517275665487562, "rewards/combined_reward/mean": 8.258497374398369, "rewards/combined_reward/std": 2.2757099526269093, "step": 26820 }, { "clip_ratio/high_max": 0.0022520947532029822, "clip_ratio/high_mean": 0.0012901346111902968, "clip_ratio/low_mean": 0.0008910319746064487, "clip_ratio/low_min": 0.00026577140354978234, "clip_ratio/region_mean": 0.0021811666054418312, "completions/clipped_ratio": 0.0, "completions/max_length": 554.4, "completions/max_terminated_length": 554.4, "completions/mean_length": 199.00834197998046, "completions/mean_terminated_length": 199.00834197998046, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.0353625755601873, "epoch": 16.432200795837158, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.0072021484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2749038352.0, "reward": 8.12613615989685, "reward_std": 0.3442035004496574, "rewards/combined_reward/mean": 8.126135635375977, "rewards/combined_reward/std": 2.35350501537323, "step": 26850 }, { "clip_ratio/high_max": 0.0016141157985354463, "clip_ratio/high_mean": 0.0009408644255017862, "clip_ratio/low_mean": 0.0007005150967112665, "clip_ratio/low_min": 0.0002434097851316134, "clip_ratio/region_mean": 0.0016413795126330418, "completions/clipped_ratio": 0.0, "completions/max_length": 614.3, "completions/max_terminated_length": 614.3, "completions/mean_length": 192.0054733276367, "completions/mean_terminated_length": 192.0054733276367, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.02753743004674713, "epoch": 16.450566268748087, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2752075813.0, "reward": 8.080236530303955, "reward_std": 0.3308753356337547, "rewards/combined_reward/mean": 8.080236148834228, "rewards/combined_reward/std": 2.220945858955383, "step": 26880 }, { "clip_ratio/high_max": 0.0018636986943117032, "clip_ratio/high_mean": 0.0010228114584606373, "clip_ratio/low_mean": 0.0007867089723731624, "clip_ratio/low_min": 0.00023298935072186094, "clip_ratio/region_mean": 0.0018095204412626722, "completions/clipped_ratio": 0.0, "completions/max_length": 677.6, "completions/max_terminated_length": 677.6, "completions/mean_length": 213.0562545776367, "completions/mean_terminated_length": 213.0562545776367, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.030207744349415103, "epoch": 16.468931741659013, "frac_reward_zero_std": 0.13333333507180214, "grad_norm": 0.006988525390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2755194109.0, "reward": 7.947141838073731, "reward_std": 0.3139538481831551, "rewards/combined_reward/mean": 7.9471415996551515, "rewards/combined_reward/std": 2.3465951681137085, "step": 26910 }, { "clip_ratio/high_max": 0.001782256025277699, "clip_ratio/high_mean": 0.00098213785495318, "clip_ratio/low_mean": 0.0007663444310310296, "clip_ratio/low_min": 0.00022663639659488885, "clip_ratio/region_mean": 0.001748482290713582, "completions/clipped_ratio": 0.0, "completions/max_length": 541.8, "completions/max_terminated_length": 541.8, "completions/mean_length": 194.01172485351563, "completions/mean_terminated_length": 194.01172485351563, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02640758970131477, "epoch": 16.487297214569942, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2758239274.0, "reward": 8.053012371063232, "reward_std": 0.3179372727870941, "rewards/combined_reward/mean": 8.05301194190979, "rewards/combined_reward/std": 2.3629958271980285, "step": 26940 }, { "clip_ratio/high_max": 0.002106677632643065, "clip_ratio/high_mean": 0.0011675014070836672, "clip_ratio/low_mean": 0.000686443799471211, "clip_ratio/low_min": 0.00018622362076712307, "clip_ratio/region_mean": 0.001853945201340442, "completions/clipped_ratio": 0.0, "completions/max_length": 554.8, "completions/max_terminated_length": 554.8, "completions/mean_length": 200.51693267822264, "completions/mean_terminated_length": 200.51693267822264, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.02875946693432828, "epoch": 16.505662687480868, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0274658203125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2761309419.0, "reward": 8.085079097747803, "reward_std": 0.35853505209088327, "rewards/combined_reward/mean": 8.085079002380372, "rewards/combined_reward/std": 2.1154677987098696, "step": 26970 }, { "clip_ratio/high_max": 0.001537830914700559, "clip_ratio/high_mean": 0.0008646916679329782, "clip_ratio/low_mean": 0.0005796508832039156, "clip_ratio/low_min": 0.0001410178735871644, "clip_ratio/region_mean": 0.0014443425367062445, "completions/clipped_ratio": 0.0, "completions/max_length": 549.1, "completions/max_terminated_length": 549.1, "completions/mean_length": 190.74662170410156, "completions/mean_terminated_length": 190.74662170410156, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.025239877387260398, "epoch": 16.524028160391797, "frac_reward_zero_std": 0.1833333384245634, "grad_norm": 0.017822265625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2764342046.0, "reward": 8.044418430328369, "reward_std": 0.3244677260518074, "rewards/combined_reward/mean": 8.044418144226075, "rewards/combined_reward/std": 2.2814220666885374, "step": 27000 }, { "clip_ratio/high_max": 0.0024966291991101266, "clip_ratio/high_mean": 0.0014393982753972522, "clip_ratio/low_mean": 0.0007103141853804118, "clip_ratio/low_min": 0.00021685634904618686, "clip_ratio/region_mean": 0.002149712448590435, "completions/clipped_ratio": 0.0, "completions/max_length": 536.2, "completions/max_terminated_length": 536.2, "completions/mean_length": 198.71614837646484, "completions/mean_terminated_length": 198.71614837646484, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.031165178523709376, "epoch": 16.542393633302723, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2767405276.0, "reward": 8.041233253479003, "reward_std": 0.36136031299829485, "rewards/combined_reward/mean": 8.041233158111572, "rewards/combined_reward/std": 2.3411792397499083, "step": 27030 }, { "clip_ratio/high_max": 0.0018625619006343186, "clip_ratio/high_mean": 0.0010809859328825648, "clip_ratio/low_mean": 0.0007665467395175559, "clip_ratio/low_min": 0.00022978551787673497, "clip_ratio/region_mean": 0.0018475326854968444, "completions/clipped_ratio": 0.0, "completions/max_length": 636.8, "completions/max_terminated_length": 636.8, "completions/mean_length": 204.81588897705078, "completions/mean_terminated_length": 204.81588897705078, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027748946923141678, "epoch": 16.560759106213652, "frac_reward_zero_std": 0.1916666693985462, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2770491929.0, "reward": 7.876669645309448, "reward_std": 0.34821787625551226, "rewards/combined_reward/mean": 7.876669311523438, "rewards/combined_reward/std": 2.3351453065872194, "step": 27060 }, { "clip_ratio/high_max": 0.002249307015154045, "clip_ratio/high_mean": 0.0013602371496138706, "clip_ratio/low_mean": 0.0007927884726086631, "clip_ratio/low_min": 0.0002268950549478177, "clip_ratio/region_mean": 0.0021530256155529058, "completions/clipped_ratio": 0.0, "completions/max_length": 613.7, "completions/max_terminated_length": 613.7, "completions/mean_length": 211.4742233276367, "completions/mean_terminated_length": 211.4742233276367, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.034345019080986576, "epoch": 16.579124579124578, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2773604150.0, "reward": 8.19662857055664, "reward_std": 0.3199916407465935, "rewards/combined_reward/mean": 8.196628284454345, "rewards/combined_reward/std": 2.1433865070343017, "step": 27090 }, { "clip_ratio/high_max": 0.002124527000220648, "clip_ratio/high_mean": 0.0011866044638736638, "clip_ratio/low_mean": 0.0007578039539415234, "clip_ratio/low_min": 0.00022605609459181626, "clip_ratio/region_mean": 0.0019444084347924218, "completions/clipped_ratio": 0.0, "completions/max_length": 557.9, "completions/max_terminated_length": 557.9, "completions/mean_length": 205.14401550292968, "completions/mean_terminated_length": 205.14401550292968, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028858812355125944, "epoch": 16.597490052035507, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2776692063.0, "reward": 8.158667659759521, "reward_std": 0.37791348397731783, "rewards/combined_reward/mean": 8.158667278289794, "rewards/combined_reward/std": 2.2104041934013368, "step": 27120 }, { "clip_ratio/high_max": 0.0022576282741889977, "clip_ratio/high_mean": 0.0013362714433848546, "clip_ratio/low_mean": 0.0009054567972877219, "clip_ratio/low_min": 0.0002534115895590124, "clip_ratio/region_mean": 0.002241728218117108, "completions/clipped_ratio": 0.0, "completions/max_length": 607.1, "completions/max_terminated_length": 607.1, "completions/mean_length": 189.90052642822266, "completions/mean_terminated_length": 189.90052642822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.032855768237883845, "epoch": 16.615855524946433, "frac_reward_zero_std": 0.21250000409781933, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2779721441.0, "reward": 7.84225435256958, "reward_std": 0.3283109277486801, "rewards/combined_reward/mean": 7.842254257202148, "rewards/combined_reward/std": 2.431156814098358, "step": 27150 }, { "clip_ratio/high_max": 0.0017976374147110619, "clip_ratio/high_mean": 0.001100326435941194, "clip_ratio/low_mean": 0.0006773461111151847, "clip_ratio/low_min": 0.0002199588151900874, "clip_ratio/region_mean": 0.0017776725443885274, "completions/clipped_ratio": 0.0, "completions/max_length": 480.7, "completions/max_terminated_length": 480.7, "completions/mean_length": 200.1057327270508, "completions/mean_terminated_length": 200.1057327270508, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 0.02831402636754016, "epoch": 16.634220997857362, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2782790007.0, "reward": 7.956113576889038, "reward_std": 0.34498453736305235, "rewards/combined_reward/mean": 7.956113290786743, "rewards/combined_reward/std": 2.2588088512420654, "step": 27180 }, { "clip_ratio/high_max": 0.001789152333609915, "clip_ratio/high_mean": 0.0010420432988515434, "clip_ratio/low_mean": 0.0007156394508153123, "clip_ratio/low_min": 0.0002606624545781718, "clip_ratio/region_mean": 0.0017576827568215473, "completions/clipped_ratio": 0.0, "completions/max_length": 551.8, "completions/max_terminated_length": 551.8, "completions/mean_length": 196.48515930175782, "completions/mean_terminated_length": 196.48515930175782, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.025850521322960656, "epoch": 16.652586470768288, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2785844670.0, "reward": 8.179635763168335, "reward_std": 0.321227502822876, "rewards/combined_reward/mean": 8.179635429382325, "rewards/combined_reward/std": 2.1992848992347716, "step": 27210 }, { "clip_ratio/high_max": 0.0019301164303518212, "clip_ratio/high_mean": 0.0011878707494664316, "clip_ratio/low_mean": 0.000690865603974089, "clip_ratio/low_min": 0.00024013576015325573, "clip_ratio/region_mean": 0.0018787363621716698, "completions/clipped_ratio": 0.0, "completions/max_length": 653.9, "completions/max_terminated_length": 653.9, "completions/mean_length": 189.29766235351562, "completions/mean_terminated_length": 189.29766235351562, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.028740907561344405, "epoch": 16.670951943679217, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2788871733.0, "reward": 7.939040040969848, "reward_std": 0.3405891418457031, "rewards/combined_reward/mean": 7.939039850234986, "rewards/combined_reward/std": 2.392302227020264, "step": 27240 }, { "clip_ratio/high_max": 0.001769800366795001, "clip_ratio/high_mean": 0.000898444317378259, "clip_ratio/low_mean": 0.0006247060549261126, "clip_ratio/low_min": 0.0001581147031780953, "clip_ratio/region_mean": 0.0015231503726681694, "completions/clipped_ratio": 0.0, "completions/max_length": 544.2, "completions/max_terminated_length": 544.2, "completions/mean_length": 182.25834045410156, "completions/mean_terminated_length": 182.25834045410156, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.028085046742732326, "epoch": 16.689317416590143, "frac_reward_zero_std": 0.20000000223517417, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2791871765.0, "reward": 8.061836051940919, "reward_std": 0.296357162296772, "rewards/combined_reward/mean": 8.061835765838623, "rewards/combined_reward/std": 2.2777783155441282, "step": 27270 }, { "clip_ratio/high_max": 0.001910524417568619, "clip_ratio/high_mean": 0.0010926809518423397, "clip_ratio/low_mean": 0.0007833930518245325, "clip_ratio/low_min": 0.00027648591979717216, "clip_ratio/region_mean": 0.0018760740145808085, "completions/clipped_ratio": 0.0, "completions/max_length": 587.6, "completions/max_terminated_length": 587.6, "completions/mean_length": 202.82760620117188, "completions/mean_terminated_length": 202.82760620117188, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.03074632940503458, "epoch": 16.707682889501072, "frac_reward_zero_std": 0.1625000022351742, "grad_norm": 0.002685546875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2794950783.0, "reward": 8.129117870330811, "reward_std": 0.32645452618598936, "rewards/combined_reward/mean": 8.1291175365448, "rewards/combined_reward/std": 2.343991982936859, "step": 27300 }, { "clip_ratio/high_max": 0.0019202649103438791, "clip_ratio/high_mean": 0.0011429618583254826, "clip_ratio/low_mean": 0.0007494729174140957, "clip_ratio/low_min": 0.00017526436798410335, "clip_ratio/region_mean": 0.0018924347682817218, "completions/clipped_ratio": 0.0, "completions/max_length": 704.3, "completions/max_terminated_length": 704.3, "completions/mean_length": 199.0315155029297, "completions/mean_terminated_length": 199.0315155029297, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.02980531871629258, "epoch": 16.726048362411998, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.00653076171875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2798015224.0, "reward": 7.892377328872681, "reward_std": 0.3304109051823616, "rewards/combined_reward/mean": 7.892377185821533, "rewards/combined_reward/std": 2.4429004430770873, "step": 27330 }, { "clip_ratio/high_max": 0.0018305449761101046, "clip_ratio/high_mean": 0.0010379826036417703, "clip_ratio/low_mean": 0.0006683269347073898, "clip_ratio/low_min": 0.0001756593939110947, "clip_ratio/region_mean": 0.0017063095501119582, "completions/clipped_ratio": 0.0, "completions/max_length": 578.6, "completions/max_terminated_length": 578.6, "completions/mean_length": 197.1369842529297, "completions/mean_terminated_length": 197.1369842529297, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02505689359580477, "epoch": 16.744413835322927, "frac_reward_zero_std": 0.1958333406597376, "grad_norm": 0.0067138671875, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 2801072390.0, "reward": 8.177050065994262, "reward_std": 0.22878083139657973, "rewards/combined_reward/mean": 8.17704963684082, "rewards/combined_reward/std": 2.031432867050171, "step": 27360 }, { "clip_ratio/high_max": 0.0020829317373378823, "clip_ratio/high_mean": 0.0011139654938839764, "clip_ratio/low_mean": 0.0009271226602625878, "clip_ratio/low_min": 0.0002969377676587707, "clip_ratio/region_mean": 0.002041088178521022, "completions/clipped_ratio": 0.0, "completions/max_length": 580.2, "completions/max_terminated_length": 580.2, "completions/mean_length": 201.8783920288086, "completions/mean_terminated_length": 201.8783920288086, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03172161534118156, "epoch": 16.762779308233853, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.005645751953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2804147763.0, "reward": 8.202505350112915, "reward_std": 0.25402879863977434, "rewards/combined_reward/mean": 8.202504968643188, "rewards/combined_reward/std": 2.0828073143959047, "step": 27390 }, { "clip_ratio/high_max": 0.0018797937921287181, "clip_ratio/high_mean": 0.0010573414986235245, "clip_ratio/low_mean": 0.000743336491965844, "clip_ratio/low_min": 0.00014391225777217188, "clip_ratio/region_mean": 0.0018006779879215173, "completions/clipped_ratio": 0.0, "completions/max_length": 557.2, "completions/max_terminated_length": 557.2, "completions/mean_length": 184.7307357788086, "completions/mean_terminated_length": 184.7307357788086, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.027172083283464113, "epoch": 16.781144781144782, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.0030670166015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2807157289.0, "reward": 8.36399540901184, "reward_std": 0.3216810330748558, "rewards/combined_reward/mean": 8.363995218276978, "rewards/combined_reward/std": 2.0399770975112914, "step": 27420 }, { "clip_ratio/high_max": 0.002143253107108952, "clip_ratio/high_mean": 0.0011742759215849219, "clip_ratio/low_mean": 0.0007281427958029477, "clip_ratio/low_min": 0.00020053310799994506, "clip_ratio/region_mean": 0.001902418716539008, "completions/clipped_ratio": 0.0, "completions/max_length": 614.9, "completions/max_terminated_length": 614.9, "completions/mean_length": 205.3742248535156, "completions/mean_terminated_length": 205.3742248535156, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.029480215301737188, "epoch": 16.799510254055708, "frac_reward_zero_std": 0.12916666977107524, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2810246086.0, "reward": 8.086469984054565, "reward_std": 0.3327764943242073, "rewards/combined_reward/mean": 8.086469745635986, "rewards/combined_reward/std": 2.2670234084129333, "step": 27450 }, { "clip_ratio/high_max": 0.001810187286173459, "clip_ratio/high_mean": 0.0010047766209027031, "clip_ratio/low_mean": 0.0009329901564342436, "clip_ratio/low_min": 0.0002994066700921394, "clip_ratio/region_mean": 0.0019377667684845315, "completions/clipped_ratio": 0.0, "completions/max_length": 530.5, "completions/max_terminated_length": 530.5, "completions/mean_length": 200.78360137939453, "completions/mean_terminated_length": 200.78360137939453, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02750703163910657, "epoch": 16.817875726966637, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2813317255.0, "reward": 8.195828056335449, "reward_std": 0.278370563685894, "rewards/combined_reward/mean": 8.195827913284301, "rewards/combined_reward/std": 2.152826654911041, "step": 27480 }, { "clip_ratio/high_max": 0.0023532212146771298, "clip_ratio/high_mean": 0.001206123671242191, "clip_ratio/low_mean": 0.0007516195116598586, "clip_ratio/low_min": 0.00024202833543919647, "clip_ratio/region_mean": 0.001957743167925704, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 195.38021392822264, "completions/mean_terminated_length": 195.38021392822264, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.030469563913842042, "epoch": 16.836241199877563, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2816367675.0, "reward": 7.946100854873658, "reward_std": 0.3561545334756374, "rewards/combined_reward/mean": 7.946100568771362, "rewards/combined_reward/std": 2.2343427538871765, "step": 27510 }, { "clip_ratio/high_max": 0.0020560609152501757, "clip_ratio/high_mean": 0.0011643243666185298, "clip_ratio/low_mean": 0.0007527012342810243, "clip_ratio/low_min": 0.00022124997340142727, "clip_ratio/region_mean": 0.0019170256108433628, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 197.89688110351562, "completions/mean_terminated_length": 197.89688110351562, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027646688596966364, "epoch": 16.854606672788492, "frac_reward_zero_std": 0.21666667610406876, "grad_norm": 0.0264892578125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2819427759.0, "reward": 8.242661285400391, "reward_std": 0.3212721191346645, "rewards/combined_reward/mean": 8.24266095161438, "rewards/combined_reward/std": 2.0816460609436036, "step": 27540 }, { "clip_ratio/high_max": 0.001809487575762129, "clip_ratio/high_mean": 0.001110686557998027, "clip_ratio/low_mean": 0.0006672695943658861, "clip_ratio/low_min": 0.0001892288938203516, "clip_ratio/region_mean": 0.0017779561448454237, "completions/clipped_ratio": 0.0, "completions/max_length": 621.4, "completions/max_terminated_length": 621.4, "completions/mean_length": 197.62838897705078, "completions/mean_terminated_length": 197.62838897705078, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03158819430197279, "epoch": 16.872972145699418, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.01190185546875, "learning_rate": 5e-06, "loss": 0.0027, "num_tokens": 2822486812.0, "reward": 8.287846612930299, "reward_std": 0.36309748888015747, "rewards/combined_reward/mean": 8.287846517562866, "rewards/combined_reward/std": 2.0277708530426026, "step": 27570 }, { "clip_ratio/high_max": 0.0020958353600387153, "clip_ratio/high_mean": 0.0011877236834455592, "clip_ratio/low_mean": 0.0008041225085131979, "clip_ratio/low_min": 0.00022924449076526797, "clip_ratio/region_mean": 0.00199184618347014, "completions/clipped_ratio": 0.0, "completions/max_length": 529.9, "completions/max_terminated_length": 529.9, "completions/mean_length": 193.38724365234376, "completions/mean_terminated_length": 193.38724365234376, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.030150318114707867, "epoch": 16.891337618610347, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.033203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2825529579.0, "reward": 8.002620220184326, "reward_std": 0.38272702097892763, "rewards/combined_reward/mean": 8.00261993408203, "rewards/combined_reward/std": 2.3885682344436647, "step": 27600 }, { "clip_ratio/high_max": 0.0017709374675177968, "clip_ratio/high_mean": 0.001050195035835107, "clip_ratio/low_mean": 0.0007488252814558412, "clip_ratio/low_min": 0.00017100774784921667, "clip_ratio/region_mean": 0.0017990203191099378, "completions/clipped_ratio": 0.0, "completions/max_length": 631.4, "completions/max_terminated_length": 631.4, "completions/mean_length": 201.95156860351562, "completions/mean_terminated_length": 201.95156860351562, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029944437881931664, "epoch": 16.909703091521273, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.00885009765625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2828605233.0, "reward": 8.16920337677002, "reward_std": 0.335803809762001, "rewards/combined_reward/mean": 8.16920280456543, "rewards/combined_reward/std": 2.0960143566131593, "step": 27630 }, { "clip_ratio/high_max": 0.0018273328358191065, "clip_ratio/high_mean": 0.0010707743140907647, "clip_ratio/low_mean": 0.0007768958347393588, "clip_ratio/low_min": 0.00025068586716467205, "clip_ratio/region_mean": 0.00184767013512707, "completions/clipped_ratio": 0.0, "completions/max_length": 553.9, "completions/max_terminated_length": 553.9, "completions/mean_length": 193.23724212646485, "completions/mean_terminated_length": 193.23724212646485, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.02989616262105604, "epoch": 16.928068564432202, "frac_reward_zero_std": 0.21250000596046448, "grad_norm": 0.034912109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2831647424.0, "reward": 7.7810293674469, "reward_std": 0.350278303027153, "rewards/combined_reward/mean": 7.781029129028321, "rewards/combined_reward/std": 2.6656413197517397, "step": 27660 }, { "clip_ratio/high_max": 0.0018926773108735992, "clip_ratio/high_mean": 0.0010708703490915166, "clip_ratio/low_mean": 0.0007152481540591301, "clip_ratio/low_min": 0.00020649717553169466, "clip_ratio/region_mean": 0.0017861184930855719, "completions/clipped_ratio": 0.0, "completions/max_length": 533.7, "completions/max_terminated_length": 533.7, "completions/mean_length": 186.72500457763672, "completions/mean_terminated_length": 186.72500457763672, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02799279095294575, "epoch": 16.946434037343128, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2834664608.0, "reward": 8.062302446365356, "reward_std": 0.296698109805584, "rewards/combined_reward/mean": 8.062301969528198, "rewards/combined_reward/std": 2.150221419334412, "step": 27690 }, { "clip_ratio/high_max": 0.001987838383744626, "clip_ratio/high_mean": 0.0011496491730213164, "clip_ratio/low_mean": 0.0008277012060716515, "clip_ratio/low_min": 0.00029023663034119334, "clip_ratio/region_mean": 0.001977350377516511, "completions/clipped_ratio": 0.0, "completions/max_length": 563.8, "completions/max_terminated_length": 563.8, "completions/mean_length": 205.67708740234374, "completions/mean_terminated_length": 205.67708740234374, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.030887772360195718, "epoch": 16.964799510254057, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.00543212890625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2837754568.0, "reward": 7.8634600162506105, "reward_std": 0.3038120731711388, "rewards/combined_reward/mean": 7.863459682464599, "rewards/combined_reward/std": 2.3417487382888793, "step": 27720 }, { "clip_ratio/high_max": 0.0019239899309468455, "clip_ratio/high_mean": 0.0010997217786401356, "clip_ratio/low_mean": 0.0008464180022807947, "clip_ratio/low_min": 0.00028789952581670755, "clip_ratio/region_mean": 0.0019461397959579093, "completions/clipped_ratio": 0.0, "completions/max_length": 595.3, "completions/max_terminated_length": 595.3, "completions/mean_length": 195.22709045410156, "completions/mean_terminated_length": 195.22709045410156, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.02997279867219428, "epoch": 16.983164983164983, "frac_reward_zero_std": 0.1958333406597376, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 2840804400.0, "reward": 8.00228934288025, "reward_std": 0.29178759306669233, "rewards/combined_reward/mean": 8.00228910446167, "rewards/combined_reward/std": 2.4390799760818482, "step": 27750 }, { "clip_ratio/high_max": 0.0017735538045211589, "clip_ratio/high_mean": 0.0010212911631660077, "clip_ratio/low_mean": 0.0007336445069980624, "clip_ratio/low_min": 0.0001963480651477871, "clip_ratio/region_mean": 0.0017549356781556959, "completions/clipped_ratio": 0.0, "completions/max_length": 638.9, "completions/max_terminated_length": 638.9, "completions/mean_length": 202.37942962646486, "completions/mean_terminated_length": 202.37942962646486, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.026703125285748086, "epoch": 17.00183654729109, "frac_reward_zero_std": 0.1416666705161333, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 2843881697.0, "reward": 8.089418506622314, "reward_std": 0.3076585680246353, "rewards/combined_reward/mean": 8.089417934417725, "rewards/combined_reward/std": 2.175626838207245, "step": 27780 }, { "clip_ratio/high_max": 0.002788112617175405, "clip_ratio/high_mean": 0.001500849883935492, "clip_ratio/low_mean": 0.0008850324196828296, "clip_ratio/low_min": 0.00023634743095802453, "clip_ratio/region_mean": 0.002385882307135034, "completions/clipped_ratio": 0.0, "completions/max_length": 605.7, "completions/max_terminated_length": 605.7, "completions/mean_length": 200.5690170288086, "completions/mean_terminated_length": 200.5690170288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02968241156389316, "epoch": 17.02020202020202, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.025146484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2846952042.0, "reward": 7.893484258651734, "reward_std": 0.3389284163713455, "rewards/combined_reward/mean": 7.893483829498291, "rewards/combined_reward/std": 2.2190981030464174, "step": 27810 }, { "clip_ratio/high_max": 0.002619350391129653, "clip_ratio/high_mean": 0.0013697261708633353, "clip_ratio/low_mean": 0.00104150740832362, "clip_ratio/low_min": 0.00034112060044814524, "clip_ratio/region_mean": 0.0024112335789444237, "completions/clipped_ratio": 0.0, "completions/max_length": 549.5, "completions/max_terminated_length": 549.5, "completions/mean_length": 207.70313110351563, "completions/mean_terminated_length": 207.70313110351563, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.031287251599133016, "epoch": 17.038567493112946, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.01385498046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2850049782.0, "reward": 8.13361678123474, "reward_std": 0.32016985714435575, "rewards/combined_reward/mean": 8.133616638183593, "rewards/combined_reward/std": 2.357408678531647, "step": 27840 }, { "clip_ratio/high_max": 0.002290448494507776, "clip_ratio/high_mean": 0.001217435581687217, "clip_ratio/low_mean": 0.0009571265417131751, "clip_ratio/low_min": 0.0002149030956691907, "clip_ratio/region_mean": 0.0021745621343143284, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 202.65599517822267, "completions/mean_terminated_length": 202.65599517822267, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027595671390493712, "epoch": 17.056932966023876, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.0303955078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2853128141.0, "reward": 7.947257995605469, "reward_std": 0.3136024706065655, "rewards/combined_reward/mean": 7.947257709503174, "rewards/combined_reward/std": 2.444755744934082, "step": 27870 }, { "clip_ratio/high_max": 0.0025486359794740566, "clip_ratio/high_mean": 0.0014773967955989064, "clip_ratio/low_mean": 0.0009916709364915731, "clip_ratio/low_min": 0.0002790729617117904, "clip_ratio/region_mean": 0.00246906773148415, "completions/clipped_ratio": 0.0, "completions/max_length": 624.1, "completions/max_terminated_length": 624.1, "completions/mean_length": 205.68463897705078, "completions/mean_terminated_length": 205.68463897705078, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.031064151351650558, "epoch": 17.0752984389348, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.003997802734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2856218130.0, "reward": 8.145962619781494, "reward_std": 0.31412080228328704, "rewards/combined_reward/mean": 8.145962381362915, "rewards/combined_reward/std": 2.2053372859954834, "step": 27900 }, { "clip_ratio/high_max": 0.0020245586036859702, "clip_ratio/high_mean": 0.0010619924102987473, "clip_ratio/low_mean": 0.0009659099036070984, "clip_ratio/low_min": 0.0002686639660775351, "clip_ratio/region_mean": 0.002027902307357484, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 198.22005920410157, "completions/mean_terminated_length": 198.22005920410157, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.026698136834117273, "epoch": 17.09366391184573, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2859279455.0, "reward": 7.899898099899292, "reward_std": 0.29836333021521566, "rewards/combined_reward/mean": 7.899897909164428, "rewards/combined_reward/std": 2.4921501755714415, "step": 27930 }, { "clip_ratio/high_max": 0.0023415043503822136, "clip_ratio/high_mean": 0.0013506544778162303, "clip_ratio/low_mean": 0.0009509918294497765, "clip_ratio/low_min": 0.0002685435147820196, "clip_ratio/region_mean": 0.002301646285923198, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 193.3885467529297, "completions/mean_terminated_length": 193.3885467529297, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02693145568482578, "epoch": 17.112029384756656, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2862322227.0, "reward": 8.245512199401855, "reward_std": 0.37170042395591735, "rewards/combined_reward/mean": 8.245512056350709, "rewards/combined_reward/std": 2.051708710193634, "step": 27960 }, { "clip_ratio/high_max": 0.0018737151733754823, "clip_ratio/high_mean": 0.0010532430436190527, "clip_ratio/low_mean": 0.0009688674908829853, "clip_ratio/low_min": 0.00031635120685677973, "clip_ratio/region_mean": 0.0020221105243156973, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 212.00755920410157, "completions/mean_terminated_length": 212.00755920410157, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.025764548142130177, "epoch": 17.130394857667586, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2865436496.0, "reward": 8.329000186920165, "reward_std": 0.32047155648469927, "rewards/combined_reward/mean": 8.328999662399292, "rewards/combined_reward/std": 2.0115397214889525, "step": 27990 }, { "clip_ratio/high_max": 0.0030082672361459117, "clip_ratio/high_mean": 0.001595675671827242, "clip_ratio/low_mean": 0.0008437236798878681, "clip_ratio/low_min": 0.00018740795057965444, "clip_ratio/region_mean": 0.0024393993623865145, "completions/clipped_ratio": 0.0, "completions/max_length": 700.2, "completions/max_terminated_length": 700.2, "completions/mean_length": 196.42969207763673, "completions/mean_terminated_length": 196.42969207763673, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03210299574614813, "epoch": 17.14876033057851, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2868490946.0, "reward": 8.063413858413696, "reward_std": 0.3328655555844307, "rewards/combined_reward/mean": 8.063413619995117, "rewards/combined_reward/std": 2.187313234806061, "step": 28020 }, { "clip_ratio/high_max": 0.0028605702435015702, "clip_ratio/high_mean": 0.0017187209023783604, "clip_ratio/low_mean": 0.0011093759484235002, "clip_ratio/low_min": 0.00031595488641566284, "clip_ratio/region_mean": 0.002828096840918685, "completions/clipped_ratio": 0.0, "completions/max_length": 603.1, "completions/max_terminated_length": 603.1, "completions/mean_length": 195.36693115234374, "completions/mean_terminated_length": 195.36693115234374, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03491427305464943, "epoch": 17.16712580348944, "frac_reward_zero_std": 0.2083333358168602, "grad_norm": 0.007293701171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2871541315.0, "reward": 7.9466992855072025, "reward_std": 0.30106047689914706, "rewards/combined_reward/mean": 7.946699142456055, "rewards/combined_reward/std": 2.3657312989234924, "step": 28050 }, { "clip_ratio/high_max": 0.0023710005606214207, "clip_ratio/high_mean": 0.001370732212914542, "clip_ratio/low_mean": 0.0010459220146003644, "clip_ratio/low_min": 0.00022785151450079865, "clip_ratio/region_mean": 0.0024166542469174603, "completions/clipped_ratio": 0.0, "completions/max_length": 568.7, "completions/max_terminated_length": 568.7, "completions/mean_length": 182.09740142822267, "completions/mean_terminated_length": 182.09740142822267, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.029356483990947407, "epoch": 17.185491276400366, "frac_reward_zero_std": 0.23750000745058059, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2874540729.0, "reward": 7.790756607055664, "reward_std": 0.2879134178161621, "rewards/combined_reward/mean": 7.790756273269653, "rewards/combined_reward/std": 2.6198875427246096, "step": 28080 }, { "clip_ratio/high_max": 0.0023553545058045224, "clip_ratio/high_mean": 0.001193738210107161, "clip_ratio/low_mean": 0.0008391768125875388, "clip_ratio/low_min": 0.00019562462888037164, "clip_ratio/region_mean": 0.002032915025968881, "completions/clipped_ratio": 0.0, "completions/max_length": 611.8, "completions/max_terminated_length": 611.8, "completions/mean_length": 200.8580780029297, "completions/mean_terminated_length": 200.8580780029297, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.027188312836612265, "epoch": 17.203856749311296, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.0264892578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2877612184.0, "reward": 8.029831647872925, "reward_std": 0.3527464345097542, "rewards/combined_reward/mean": 8.029831314086914, "rewards/combined_reward/std": 2.413124752044678, "step": 28110 }, { "clip_ratio/high_max": 0.002558662192313932, "clip_ratio/high_mean": 0.001448217223029739, "clip_ratio/low_mean": 0.0007879985977221319, "clip_ratio/low_min": 0.00021454568219875608, "clip_ratio/region_mean": 0.002236215817295791, "completions/clipped_ratio": 0.0, "completions/max_length": 508.8, "completions/max_terminated_length": 508.8, "completions/mean_length": 202.24323577880858, "completions/mean_terminated_length": 202.24323577880858, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028473489234844843, "epoch": 17.22222222222222, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2880688958.0, "reward": 8.048625612258911, "reward_std": 0.25000266432762147, "rewards/combined_reward/mean": 8.048625040054322, "rewards/combined_reward/std": 2.322134220600128, "step": 28140 }, { "clip_ratio/high_max": 0.0021216647350229324, "clip_ratio/high_mean": 0.0012297598327374241, "clip_ratio/low_mean": 0.001006627714135296, "clip_ratio/low_min": 0.0003303432660080337, "clip_ratio/region_mean": 0.002236387553663614, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 198.4250061035156, "completions/mean_terminated_length": 198.4250061035156, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027704849699512125, "epoch": 17.24058769513315, "frac_reward_zero_std": 0.19166667014360428, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2883751070.0, "reward": 8.11734914779663, "reward_std": 0.3075664281845093, "rewards/combined_reward/mean": 8.117348861694335, "rewards/combined_reward/std": 2.152225971221924, "step": 28170 }, { "clip_ratio/high_max": 0.0022926082281628625, "clip_ratio/high_mean": 0.0014102525305740225, "clip_ratio/low_mean": 0.0009825682792628263, "clip_ratio/low_min": 0.0002898308402412416, "clip_ratio/region_mean": 0.0023928207888578375, "completions/clipped_ratio": 0.0, "completions/max_length": 560.4, "completions/max_terminated_length": 560.4, "completions/mean_length": 192.74297332763672, "completions/mean_terminated_length": 192.74297332763672, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.029643167943383255, "epoch": 17.258953168044076, "frac_reward_zero_std": 0.16250000186264515, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2886791363.0, "reward": 8.127969455718993, "reward_std": 0.2713300667703152, "rewards/combined_reward/mean": 8.127969217300414, "rewards/combined_reward/std": 2.1323394536972047, "step": 28200 }, { "clip_ratio/high_max": 0.0021714404327212835, "clip_ratio/high_mean": 0.0011891323602564322, "clip_ratio/low_mean": 0.00097224362786316, "clip_ratio/low_min": 0.0002800694286027768, "clip_ratio/region_mean": 0.002161376004611763, "completions/clipped_ratio": 0.0, "completions/max_length": 592.6, "completions/max_terminated_length": 592.6, "completions/mean_length": 202.9510467529297, "completions/mean_terminated_length": 202.9510467529297, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.02723058994549016, "epoch": 17.277318640955006, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2889870855.0, "reward": 8.276213264465332, "reward_std": 0.34493753015995027, "rewards/combined_reward/mean": 8.276213026046753, "rewards/combined_reward/std": 2.0330080509185793, "step": 28230 }, { "clip_ratio/high_max": 0.0022738020488759503, "clip_ratio/high_mean": 0.0013246645878098207, "clip_ratio/low_mean": 0.0008502619471376723, "clip_ratio/low_min": 0.0001345063152257353, "clip_ratio/region_mean": 0.0021749265356144558, "completions/clipped_ratio": 0.0, "completions/max_length": 516.4, "completions/max_terminated_length": 516.4, "completions/mean_length": 195.1164093017578, "completions/mean_terminated_length": 195.1164093017578, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.02630622571644684, "epoch": 17.29568411386593, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.003753662109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2892920262.0, "reward": 8.296230506896972, "reward_std": 0.3278196111321449, "rewards/combined_reward/mean": 8.296229934692382, "rewards/combined_reward/std": 1.957448673248291, "step": 28260 }, { "clip_ratio/high_max": 0.0021062405304595206, "clip_ratio/high_mean": 0.001166013002997109, "clip_ratio/low_mean": 0.0008096395507891429, "clip_ratio/low_min": 0.00016595099441474304, "clip_ratio/region_mean": 0.0019756525585156244, "completions/clipped_ratio": 0.0, "completions/max_length": 607.7, "completions/max_terminated_length": 607.7, "completions/mean_length": 198.38021392822264, "completions/mean_terminated_length": 198.38021392822264, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.027246665217292805, "epoch": 17.31404958677686, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.006744384765625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2895982202.0, "reward": 8.113224172592163, "reward_std": 0.2894704908132553, "rewards/combined_reward/mean": 8.113224124908447, "rewards/combined_reward/std": 2.0926266074180604, "step": 28290 }, { "clip_ratio/high_max": 0.0028649891333770937, "clip_ratio/high_mean": 0.0015269429618153178, "clip_ratio/low_mean": 0.0007986334781890037, "clip_ratio/low_min": 0.00021500477475152973, "clip_ratio/region_mean": 0.00232557644485496, "completions/clipped_ratio": 0.0, "completions/max_length": 562.2, "completions/max_terminated_length": 562.2, "completions/mean_length": 203.5377670288086, "completions/mean_terminated_length": 203.5377670288086, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.03025151204007367, "epoch": 17.332415059687786, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.0260009765625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2899063947.0, "reward": 7.990376567840576, "reward_std": 0.3186698704957962, "rewards/combined_reward/mean": 7.990376567840576, "rewards/combined_reward/std": 2.4971876740455627, "step": 28320 }, { "clip_ratio/high_max": 0.0024222534586442633, "clip_ratio/high_mean": 0.0012757300115481484, "clip_ratio/low_mean": 0.0009198785253829555, "clip_ratio/low_min": 0.000251799516020886, "clip_ratio/region_mean": 0.002195608558637711, "completions/clipped_ratio": 0.0, "completions/max_length": 517.9, "completions/max_terminated_length": 517.9, "completions/mean_length": 195.8182357788086, "completions/mean_terminated_length": 195.8182357788086, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.0262976232295235, "epoch": 17.350780532598716, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.00567626953125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 2902116049.0, "reward": 7.980292701721192, "reward_std": 0.2587678015232086, "rewards/combined_reward/mean": 7.9802923679351805, "rewards/combined_reward/std": 2.361038875579834, "step": 28350 }, { "clip_ratio/high_max": 0.002384891861098974, "clip_ratio/high_mean": 0.0013946965540526436, "clip_ratio/low_mean": 0.0010159275033705248, "clip_ratio/low_min": 0.00030958706823488075, "clip_ratio/region_mean": 0.0024106240443264445, "completions/clipped_ratio": 0.0, "completions/max_length": 610.7, "completions/max_terminated_length": 610.7, "completions/mean_length": 202.48385925292968, "completions/mean_terminated_length": 202.48385925292968, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.03220612839795649, "epoch": 17.36914600550964, "frac_reward_zero_std": 0.15000000186264514, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2905193747.0, "reward": 8.297109222412109, "reward_std": 0.35821979865431786, "rewards/combined_reward/mean": 8.297108888626099, "rewards/combined_reward/std": 2.006161105632782, "step": 28380 }, { "clip_ratio/high_max": 0.0022258485570394743, "clip_ratio/high_mean": 0.0013130178881207637, "clip_ratio/low_mean": 0.0008570507131177388, "clip_ratio/low_min": 0.00025471095674826455, "clip_ratio/region_mean": 0.0021700686066954706, "completions/clipped_ratio": 0.0, "completions/max_length": 657.1, "completions/max_terminated_length": 657.1, "completions/mean_length": 202.7697967529297, "completions/mean_terminated_length": 202.7697967529297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02761182812973857, "epoch": 17.38751147842057, "frac_reward_zero_std": 0.2125000052154064, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2908272543.0, "reward": 8.158820629119873, "reward_std": 0.33646187633275987, "rewards/combined_reward/mean": 8.158820390701294, "rewards/combined_reward/std": 2.07671457529068, "step": 28410 }, { "clip_ratio/high_max": 0.0024209582077067654, "clip_ratio/high_mean": 0.001425332488846228, "clip_ratio/low_mean": 0.0010902060975543767, "clip_ratio/low_min": 0.00032625153811144023, "clip_ratio/region_mean": 0.00251553856942337, "completions/clipped_ratio": 0.0, "completions/max_length": 691.8, "completions/max_terminated_length": 691.8, "completions/mean_length": 203.1291702270508, "completions/mean_terminated_length": 203.1291702270508, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.030036825368491313, "epoch": 17.405876951331496, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2911352719.0, "reward": 7.943912363052368, "reward_std": 0.32198126763105395, "rewards/combined_reward/mean": 7.943912029266357, "rewards/combined_reward/std": 2.3850549697875976, "step": 28440 }, { "clip_ratio/high_max": 0.002533319641467339, "clip_ratio/high_mean": 0.0014233134085467707, "clip_ratio/low_mean": 0.0008386638846180479, "clip_ratio/low_min": 0.00019775116015807725, "clip_ratio/region_mean": 0.002261977276551382, "completions/clipped_ratio": 0.0, "completions/max_length": 577.3, "completions/max_terminated_length": 577.3, "completions/mean_length": 200.47474517822266, "completions/mean_terminated_length": 200.47474517822266, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.02776847210091849, "epoch": 17.424242424242426, "frac_reward_zero_std": 0.19583334252238274, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2914422702.0, "reward": 7.922149753570556, "reward_std": 0.3799358606338501, "rewards/combined_reward/mean": 7.922149419784546, "rewards/combined_reward/std": 2.4479647517204284, "step": 28470 }, { "clip_ratio/high_max": 0.002100282153696753, "clip_ratio/high_mean": 0.0011309759288754624, "clip_ratio/low_mean": 0.0008425711103579185, "clip_ratio/low_min": 0.00019683125283336266, "clip_ratio/region_mean": 0.0019735470351103382, "completions/clipped_ratio": 0.0, "completions/max_length": 522.6, "completions/max_terminated_length": 522.6, "completions/mean_length": 183.63021087646484, "completions/mean_terminated_length": 183.63021087646484, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026840036168384054, "epoch": 17.44260789715335, "frac_reward_zero_std": 0.20833333991467953, "grad_norm": 0.0084228515625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2917428002.0, "reward": 8.276723432540894, "reward_std": 0.31262879222631457, "rewards/combined_reward/mean": 8.276722860336303, "rewards/combined_reward/std": 2.134537363052368, "step": 28500 }, { "clip_ratio/high_max": 0.002309800500127797, "clip_ratio/high_mean": 0.0013887990643221806, "clip_ratio/low_mean": 0.0010190413229186864, "clip_ratio/low_min": 0.00028213851425486306, "clip_ratio/region_mean": 0.002407840368687175, "completions/clipped_ratio": 0.0, "completions/max_length": 613.1, "completions/max_terminated_length": 613.1, "completions/mean_length": 204.79635925292968, "completions/mean_terminated_length": 204.79635925292968, "completions/min_length": 47.7, "completions/min_terminated_length": 47.7, "entropy": 0.029668199317529797, "epoch": 17.46097337006428, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2920514580.0, "reward": 8.090389156341553, "reward_std": 0.34978306442499163, "rewards/combined_reward/mean": 8.090388822555543, "rewards/combined_reward/std": 2.406837010383606, "step": 28530 }, { "clip_ratio/high_max": 0.0021836390796427926, "clip_ratio/high_mean": 0.0012710021623206559, "clip_ratio/low_mean": 0.0008987982704032523, "clip_ratio/low_min": 0.00022691077465424313, "clip_ratio/region_mean": 0.0021698004367256846, "completions/clipped_ratio": 0.0, "completions/max_length": 518.8, "completions/max_terminated_length": 518.8, "completions/mean_length": 199.4757858276367, "completions/mean_terminated_length": 199.4757858276367, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.02584939622320235, "epoch": 17.479338842975206, "frac_reward_zero_std": 0.16666666865348817, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2923580727.0, "reward": 8.0583411693573, "reward_std": 0.3208311602473259, "rewards/combined_reward/mean": 8.05834083557129, "rewards/combined_reward/std": 2.1963231325149537, "step": 28560 }, { "clip_ratio/high_max": 0.0023730546830241415, "clip_ratio/high_mean": 0.0013148907552628468, "clip_ratio/low_mean": 0.0008606726240638333, "clip_ratio/low_min": 0.00020300217171704087, "clip_ratio/region_mean": 0.0021755633657448924, "completions/clipped_ratio": 0.0, "completions/max_length": 536.5, "completions/max_terminated_length": 536.5, "completions/mean_length": 200.21276397705077, "completions/mean_terminated_length": 200.21276397705077, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02753013528417796, "epoch": 17.497704315886136, "frac_reward_zero_std": 0.19583333767950534, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2926649704.0, "reward": 8.02142391204834, "reward_std": 0.34000255465507506, "rewards/combined_reward/mean": 8.021423530578613, "rewards/combined_reward/std": 2.405216300487518, "step": 28590 }, { "clip_ratio/high_max": 0.002131958158376316, "clip_ratio/high_mean": 0.001240832041609489, "clip_ratio/low_mean": 0.0009144541122320031, "clip_ratio/low_min": 0.00018919236536021344, "clip_ratio/region_mean": 0.002155286153720226, "completions/clipped_ratio": 0.0, "completions/max_length": 521.5, "completions/max_terminated_length": 521.5, "completions/mean_length": 195.2854248046875, "completions/mean_terminated_length": 195.2854248046875, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.025917952026550967, "epoch": 17.51606978879706, "frac_reward_zero_std": 0.21666667312383653, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2929699760.0, "reward": 8.138630533218384, "reward_std": 0.26543503254652023, "rewards/combined_reward/mean": 8.13863010406494, "rewards/combined_reward/std": 2.1328552603721618, "step": 28620 }, { "clip_ratio/high_max": 0.0025039530989791575, "clip_ratio/high_mean": 0.0015182227558398153, "clip_ratio/low_mean": 0.0009078369415268147, "clip_ratio/low_min": 0.0002430437885777792, "clip_ratio/region_mean": 0.0024260596993068853, "completions/clipped_ratio": 0.0, "completions/max_length": 611.9, "completions/max_terminated_length": 611.9, "completions/mean_length": 210.30469360351563, "completions/mean_terminated_length": 210.30469360351563, "completions/min_length": 46.1, "completions/min_terminated_length": 46.1, "entropy": 0.029075166629627346, "epoch": 17.53443526170799, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0035552978515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2932807490.0, "reward": 8.042332887649536, "reward_std": 0.32404097244143487, "rewards/combined_reward/mean": 8.04233283996582, "rewards/combined_reward/std": 2.2760738372802733, "step": 28650 }, { "clip_ratio/high_max": 0.0029521089949412272, "clip_ratio/high_mean": 0.001500952775919965, "clip_ratio/low_mean": 0.0010801401622302365, "clip_ratio/low_min": 0.0003056870307773352, "clip_ratio/region_mean": 0.002581092940332989, "completions/clipped_ratio": 0.0, "completions/max_length": 633.6, "completions/max_terminated_length": 633.6, "completions/mean_length": 195.72786865234374, "completions/mean_terminated_length": 195.72786865234374, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.031691441560784975, "epoch": 17.552800734618916, "frac_reward_zero_std": 0.20000000931322576, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2935859245.0, "reward": 8.04671688079834, "reward_std": 0.28036039471626284, "rewards/combined_reward/mean": 8.04671688079834, "rewards/combined_reward/std": 2.428124725818634, "step": 28680 }, { "clip_ratio/high_max": 0.00245881255104905, "clip_ratio/high_mean": 0.0014432848638534779, "clip_ratio/low_mean": 0.001012285682008951, "clip_ratio/low_min": 0.0003210756896199503, "clip_ratio/region_mean": 0.0024555705337358327, "completions/clipped_ratio": 0.0, "completions/max_length": 511.7, "completions/max_terminated_length": 511.7, "completions/mean_length": 189.80573272705078, "completions/mean_terminated_length": 189.80573272705078, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02995490130657951, "epoch": 17.571166207529846, "frac_reward_zero_std": 0.21666667610406876, "grad_norm": 0.0279541015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2938888259.0, "reward": 8.025831651687621, "reward_std": 0.3472082644701004, "rewards/combined_reward/mean": 8.025831413269042, "rewards/combined_reward/std": 2.295971357822418, "step": 28710 }, { "clip_ratio/high_max": 0.0024400394497206433, "clip_ratio/high_mean": 0.0014030544783357375, "clip_ratio/low_mean": 0.0008960007273951002, "clip_ratio/low_min": 0.00023504441754388, "clip_ratio/region_mean": 0.0022990552184637635, "completions/clipped_ratio": 0.0, "completions/max_length": 559.6, "completions/max_terminated_length": 559.6, "completions/mean_length": 189.9026092529297, "completions/mean_terminated_length": 189.9026092529297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03003034400753677, "epoch": 17.58953168044077, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2941917645.0, "reward": 7.958383560180664, "reward_std": 0.2665646642446518, "rewards/combined_reward/mean": 7.958383512496948, "rewards/combined_reward/std": 2.5352837204933167, "step": 28740 }, { "clip_ratio/high_max": 0.0030446317621681375, "clip_ratio/high_mean": 0.0016587625861575362, "clip_ratio/low_mean": 0.0009903910500725032, "clip_ratio/low_min": 0.0003119774026951442, "clip_ratio/region_mean": 0.002649153652600944, "completions/clipped_ratio": 0.0, "completions/max_length": 589.1, "completions/max_terminated_length": 589.1, "completions/mean_length": 202.19219360351562, "completions/mean_terminated_length": 202.19219360351562, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.03322415991375844, "epoch": 17.6078971533517, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 2944994223.0, "reward": 7.993010425567627, "reward_std": 0.37876026183366773, "rewards/combined_reward/mean": 7.9930100440979, "rewards/combined_reward/std": 2.2338061928749084, "step": 28770 }, { "clip_ratio/high_max": 0.002774934207263868, "clip_ratio/high_mean": 0.0015793136136683947, "clip_ratio/low_mean": 0.0011963446494822469, "clip_ratio/low_min": 0.00028803848787598935, "clip_ratio/region_mean": 0.0027756582586638007, "completions/clipped_ratio": 0.0, "completions/max_length": 545.7, "completions/max_terminated_length": 545.7, "completions/mean_length": 189.70417022705078, "completions/mean_terminated_length": 189.70417022705078, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03336387096593777, "epoch": 17.626262626262626, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.03125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2948022847.0, "reward": 8.0596333026886, "reward_std": 0.3664874494075775, "rewards/combined_reward/mean": 8.059633159637452, "rewards/combined_reward/std": 2.0769304990768434, "step": 28800 }, { "clip_ratio/high_max": 0.002183899007407793, "clip_ratio/high_mean": 0.0011875910990056583, "clip_ratio/low_mean": 0.0007187803072156385, "clip_ratio/low_min": 0.00021715438706451095, "clip_ratio/region_mean": 0.001906371414346116, "completions/clipped_ratio": 0.0, "completions/max_length": 494.9, "completions/max_terminated_length": 494.9, "completions/mean_length": 194.73959197998047, "completions/mean_terminated_length": 194.73959197998047, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 0.02769081813748926, "epoch": 17.64401591674319, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 2951070807.0, "reward": 8.05762791633606, "reward_std": 0.34184303879737854, "rewards/combined_reward/mean": 8.057627487182618, "rewards/combined_reward/std": 2.177467942237854, "step": 28830 }, { "clip_ratio/high_max": 0.002076022151838212, "clip_ratio/high_mean": 0.0011250980165641521, "clip_ratio/low_mean": 0.0006742658770235721, "clip_ratio/low_min": 0.0002388588565130097, "clip_ratio/region_mean": 0.0017993638995297565, "completions/clipped_ratio": 0.0, "completions/max_length": 571.6, "completions/max_terminated_length": 571.6, "completions/mean_length": 197.84167022705077, "completions/mean_terminated_length": 197.84167022705077, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.027771582609663406, "epoch": 17.66238138965412, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2954130679.0, "reward": 8.170364570617675, "reward_std": 0.3718705877661705, "rewards/combined_reward/mean": 8.170364332199096, "rewards/combined_reward/std": 2.3088052749633787, "step": 28860 }, { "clip_ratio/high_max": 0.0021105506779955856, "clip_ratio/high_mean": 0.0011586862892727368, "clip_ratio/low_mean": 0.0007101861667858126, "clip_ratio/low_min": 8.812693946917231e-05, "clip_ratio/region_mean": 0.0018688724584838686, "completions/clipped_ratio": 0.0, "completions/max_length": 526.3, "completions/max_terminated_length": 526.3, "completions/mean_length": 183.27813110351562, "completions/mean_terminated_length": 183.27813110351562, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027083625877276064, "epoch": 17.680746862565044, "frac_reward_zero_std": 0.23333334177732468, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2957134627.0, "reward": 7.977370500564575, "reward_std": 0.31152852773666384, "rewards/combined_reward/mean": 7.977370357513427, "rewards/combined_reward/std": 2.252273201942444, "step": 28890 }, { "clip_ratio/high_max": 0.0018764319887850434, "clip_ratio/high_mean": 0.0010345472879028724, "clip_ratio/low_mean": 0.0008116803729838769, "clip_ratio/low_min": 0.00022482827625935897, "clip_ratio/region_mean": 0.0018462276619781429, "completions/clipped_ratio": 0.0, "completions/max_length": 532.5, "completions/max_terminated_length": 532.5, "completions/mean_length": 189.72005767822264, "completions/mean_terminated_length": 189.72005767822264, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03068641947271923, "epoch": 17.699112335475974, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2960163312.0, "reward": 8.11458067893982, "reward_std": 0.2718639962375164, "rewards/combined_reward/mean": 8.114580631256104, "rewards/combined_reward/std": 2.352557027339935, "step": 28920 }, { "clip_ratio/high_max": 0.002002288963800917, "clip_ratio/high_mean": 0.0011835885469433074, "clip_ratio/low_mean": 0.0008679934311658144, "clip_ratio/low_min": 0.0002583947488650059, "clip_ratio/region_mean": 0.0020515819938736967, "completions/clipped_ratio": 0.0, "completions/max_length": 697.5, "completions/max_terminated_length": 697.5, "completions/mean_length": 204.86068115234374, "completions/mean_terminated_length": 204.86068115234374, "completions/min_length": 46.7, "completions/min_terminated_length": 46.7, "entropy": 0.03064569310906033, "epoch": 17.7174778083869, "frac_reward_zero_std": 0.14583334065973758, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 2963250137.0, "reward": 7.9931004524230955, "reward_std": 0.32798906192183497, "rewards/combined_reward/mean": 7.993100500106811, "rewards/combined_reward/std": 2.371275532245636, "step": 28950 }, { "clip_ratio/high_max": 0.0017440796000300907, "clip_ratio/high_mean": 0.0010202672482895045, "clip_ratio/low_mean": 0.0008048231175052933, "clip_ratio/low_min": 0.00020500900015273752, "clip_ratio/region_mean": 0.0018250903701603723, "completions/clipped_ratio": 0.0, "completions/max_length": 529.3, "completions/max_terminated_length": 529.3, "completions/mean_length": 194.53203582763672, "completions/mean_terminated_length": 194.53203582763672, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.027445004026715954, "epoch": 17.735843281297825, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.00738525390625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 2966297300.0, "reward": 7.954726314544677, "reward_std": 0.24244431927800178, "rewards/combined_reward/mean": 7.954726266860962, "rewards/combined_reward/std": 2.347764754295349, "step": 28980 }, { "clip_ratio/high_max": 0.0017582371321623214, "clip_ratio/high_mean": 0.0010242610307614087, "clip_ratio/low_mean": 0.0007545756787294521, "clip_ratio/low_min": 0.00021506942891088936, "clip_ratio/region_mean": 0.0017788367064592119, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 208.99271545410156, "completions/mean_terminated_length": 208.99271545410156, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02982031248199443, "epoch": 17.754208754208754, "frac_reward_zero_std": 0.2000000037252903, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 2969399992.0, "reward": 8.29198932647705, "reward_std": 0.293797904998064, "rewards/combined_reward/mean": 8.29198923110962, "rewards/combined_reward/std": 1.9877153933048248, "step": 29010 }, { "clip_ratio/high_max": 0.0017901563430010963, "clip_ratio/high_mean": 0.0009601777100518423, "clip_ratio/low_mean": 0.0007865334049711237, "clip_ratio/low_min": 0.00014357444112344335, "clip_ratio/region_mean": 0.0017467111222989236, "completions/clipped_ratio": 0.0, "completions/max_length": 496.8, "completions/max_terminated_length": 496.8, "completions/mean_length": 183.7296936035156, "completions/mean_terminated_length": 183.7296936035156, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026045644120313228, "epoch": 17.77257422711968, "frac_reward_zero_std": 0.1875000014901161, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 2972405674.0, "reward": 8.259713172912598, "reward_std": 0.285669369995594, "rewards/combined_reward/mean": 8.259712982177735, "rewards/combined_reward/std": 2.074789893627167, "step": 29040 }, { "clip_ratio/high_max": 0.0020645794094889423, "clip_ratio/high_mean": 0.0011294414894412815, "clip_ratio/low_mean": 0.0007653609134043412, "clip_ratio/low_min": 0.00015822022420858655, "clip_ratio/region_mean": 0.0018948023964185268, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 200.22708892822266, "completions/mean_terminated_length": 200.22708892822266, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.03124488756681482, "epoch": 17.79093970003061, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.00714111328125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 2975474706.0, "reward": 8.058347702026367, "reward_std": 0.32883397564291955, "rewards/combined_reward/mean": 8.058347606658936, "rewards/combined_reward/std": 2.322901463508606, "step": 29070 }, { "clip_ratio/high_max": 0.002010633678097899, "clip_ratio/high_mean": 0.001173968459382498, "clip_ratio/low_mean": 0.000779540469981536, "clip_ratio/low_min": 0.00018090309701316679, "clip_ratio/region_mean": 0.0019535089629547048, "completions/clipped_ratio": 0.0, "completions/max_length": 593.8, "completions/max_terminated_length": 593.8, "completions/mean_length": 207.74505615234375, "completions/mean_terminated_length": 207.74505615234375, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.029995489725843073, "epoch": 17.809305172941535, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 2978572607.0, "reward": 8.338128471374512, "reward_std": 0.32025574147701263, "rewards/combined_reward/mean": 8.338128089904785, "rewards/combined_reward/std": 1.9866219878196716, "step": 29100 }, { "clip_ratio/high_max": 0.0021297320895731294, "clip_ratio/high_mean": 0.0011992630881043928, "clip_ratio/low_mean": 0.0008621065328043187, "clip_ratio/low_min": 0.00018025705066975205, "clip_ratio/region_mean": 0.0020613696234552965, "completions/clipped_ratio": 0.0, "completions/max_length": 519.2, "completions/max_terminated_length": 519.2, "completions/mean_length": 195.5940170288086, "completions/mean_terminated_length": 195.5940170288086, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.029625064538170893, "epoch": 17.827670645852464, "frac_reward_zero_std": 0.2083333432674408, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 2981623848.0, "reward": 8.016260766983033, "reward_std": 0.26778814420104025, "rewards/combined_reward/mean": 8.016260719299316, "rewards/combined_reward/std": 2.257879400253296, "step": 29130 }, { "clip_ratio/high_max": 0.0017858030017426547, "clip_ratio/high_mean": 0.0010668255845909394, "clip_ratio/low_mean": 0.0008025591347783726, "clip_ratio/low_min": 0.00023329695056114967, "clip_ratio/region_mean": 0.0018693846873550986, "completions/clipped_ratio": 0.0, "completions/max_length": 667.2, "completions/max_terminated_length": 667.2, "completions/mean_length": 190.02786865234376, "completions/mean_terminated_length": 190.02786865234376, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.027152511089419327, "epoch": 17.84603611876339, "frac_reward_zero_std": 0.22083333879709244, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 2984653715.0, "reward": 8.103609418869018, "reward_std": 0.3476668119430542, "rewards/combined_reward/mean": 8.103609037399291, "rewards/combined_reward/std": 2.156694400310516, "step": 29160 }, { "clip_ratio/high_max": 0.002104999358804586, "clip_ratio/high_mean": 0.0012164065254182787, "clip_ratio/low_mean": 0.0007524272537314876, "clip_ratio/low_min": 0.00020530485659643697, "clip_ratio/region_mean": 0.00196883376193, "completions/clipped_ratio": 0.0, "completions/max_length": 658.5, "completions/max_terminated_length": 658.5, "completions/mean_length": 204.42292175292968, "completions/mean_terminated_length": 204.42292175292968, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.03125098447004954, "epoch": 17.86440159167432, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.03173828125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 2987738859.0, "reward": 8.274190044403076, "reward_std": 0.32737177759408953, "rewards/combined_reward/mean": 8.274189996719361, "rewards/combined_reward/std": 2.0565188407897947, "step": 29190 }, { "clip_ratio/high_max": 0.0017679025297790455, "clip_ratio/high_mean": 0.0010436872949261063, "clip_ratio/low_mean": 0.000638624868709788, "clip_ratio/low_min": 0.00020619872084353118, "clip_ratio/region_mean": 0.0016823121589065219, "completions/clipped_ratio": 0.0, "completions/max_length": 579.9, "completions/max_terminated_length": 579.9, "completions/mean_length": 190.8619842529297, "completions/mean_terminated_length": 190.8619842529297, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028545313607901336, "epoch": 17.882767064585245, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 2990771929.0, "reward": 8.17490873336792, "reward_std": 0.37861229479312897, "rewards/combined_reward/mean": 8.17490849494934, "rewards/combined_reward/std": 2.1767019391059876, "step": 29220 }, { "clip_ratio/high_max": 0.0020998692203041476, "clip_ratio/high_mean": 0.001232643334272628, "clip_ratio/low_mean": 0.0008092343627140508, "clip_ratio/low_min": 0.0002368725836277008, "clip_ratio/region_mean": 0.0020418777135394824, "completions/clipped_ratio": 0.0, "completions/max_length": 569.9, "completions/max_terminated_length": 569.9, "completions/mean_length": 203.48984985351564, "completions/mean_terminated_length": 203.48984985351564, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.03248101593926549, "epoch": 17.901132537496174, "frac_reward_zero_std": 0.16666666828095913, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 2993853490.0, "reward": 8.053608608245849, "reward_std": 0.4026409924030304, "rewards/combined_reward/mean": 8.053608512878418, "rewards/combined_reward/std": 2.2685791730880736, "step": 29250 }, { "clip_ratio/high_max": 0.0018391366975265555, "clip_ratio/high_mean": 0.0010955944043719986, "clip_ratio/low_mean": 0.0007513842547875053, "clip_ratio/low_min": 0.0002365434234282778, "clip_ratio/region_mean": 0.0018469786446075886, "completions/clipped_ratio": 0.0, "completions/max_length": 564.5, "completions/max_terminated_length": 564.5, "completions/mean_length": 196.62474365234374, "completions/mean_terminated_length": 196.62474365234374, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.03041591450261573, "epoch": 17.9194980104071, "frac_reward_zero_std": 0.22083334177732467, "grad_norm": 0.00567626953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 2996908689.0, "reward": 8.05272822380066, "reward_std": 0.2904795601963997, "rewards/combined_reward/mean": 8.052728128433227, "rewards/combined_reward/std": 2.3967597007751467, "step": 29280 }, { "clip_ratio/high_max": 0.0017551656948247305, "clip_ratio/high_mean": 0.0010029538096811544, "clip_ratio/low_mean": 0.0008348314222530462, "clip_ratio/low_min": 0.00031044238615625847, "clip_ratio/region_mean": 0.0018377852451521904, "completions/clipped_ratio": 0.0, "completions/max_length": 536.1, "completions/max_terminated_length": 536.1, "completions/mean_length": 181.27005615234376, "completions/mean_terminated_length": 181.27005615234376, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02717412739681701, "epoch": 17.93786348331803, "frac_reward_zero_std": 0.20416667275130748, "grad_norm": 0.021728515625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 2999904926.0, "reward": 7.757593774795533, "reward_std": 0.31188881397247314, "rewards/combined_reward/mean": 7.757593441009521, "rewards/combined_reward/std": 2.5345155239105224, "step": 29310 }, { "clip_ratio/high_max": 0.002062849157664459, "clip_ratio/high_mean": 0.0011283853658824229, "clip_ratio/low_mean": 0.0008088806545856642, "clip_ratio/low_min": 0.00022650731601364288, "clip_ratio/region_mean": 0.001937266020346821, "completions/clipped_ratio": 0.0, "completions/max_length": 643.2, "completions/max_terminated_length": 643.2, "completions/mean_length": 205.37500457763673, "completions/mean_terminated_length": 205.37500457763673, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.029807579657062887, "epoch": 17.956228956228955, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3002993726.0, "reward": 8.10987572669983, "reward_std": 0.3442822530865669, "rewards/combined_reward/mean": 8.109875392913818, "rewards/combined_reward/std": 2.075299787521362, "step": 29340 }, { "clip_ratio/high_max": 0.0021407897550185834, "clip_ratio/high_mean": 0.0011854809107414135, "clip_ratio/low_mean": 0.0007292196684526667, "clip_ratio/low_min": 0.00016000779990766508, "clip_ratio/region_mean": 0.0019147005737371122, "completions/clipped_ratio": 0.0, "completions/max_length": 530.5, "completions/max_terminated_length": 530.5, "completions/mean_length": 196.75391235351563, "completions/mean_terminated_length": 196.75391235351563, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029444829265897473, "epoch": 17.974594429139884, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3006049421.0, "reward": 7.8135277271270756, "reward_std": 0.3327944248914719, "rewards/combined_reward/mean": 7.813527536392212, "rewards/combined_reward/std": 2.566271150112152, "step": 29370 }, { "clip_ratio/high_max": 0.0018476311287183003, "clip_ratio/high_mean": 0.001134526041156884, "clip_ratio/low_mean": 0.0007612216993341766, "clip_ratio/low_min": 0.00022240329126361759, "clip_ratio/region_mean": 0.0018957477412186563, "completions/clipped_ratio": 0.0, "completions/max_length": 725.5, "completions/max_terminated_length": 725.5, "completions/mean_length": 203.0713623046875, "completions/mean_terminated_length": 203.0713623046875, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03018455400597304, "epoch": 17.99295990205081, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3009129375.0, "reward": 7.946963214874268, "reward_std": 0.3405133754014969, "rewards/combined_reward/mean": 7.946962976455689, "rewards/combined_reward/std": 2.4018625378608705, "step": 29400 }, { "clip_ratio/high_max": 0.002169950591346829, "clip_ratio/high_mean": 0.0012760224182456427, "clip_ratio/low_mean": 0.0009097455907469905, "clip_ratio/low_min": 0.00028733357851065265, "clip_ratio/region_mean": 0.0021857680193698185, "completions/clipped_ratio": 0.0, "completions/max_length": 578.4545454545455, "completions/max_terminated_length": 578.4545454545455, "completions/mean_length": 203.0672413219105, "completions/mean_terminated_length": 203.0672413219105, "completions/min_length": 39.90909090909091, "completions/min_terminated_length": 39.90909090909091, "entropy": 0.030344248001204163, "epoch": 18.011631466176922, "frac_reward_zero_std": 0.16287879239429126, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3012517307.0, "reward": 8.141685442491012, "reward_std": 0.2836145826361396, "rewards/combined_reward/mean": 8.141685399142178, "rewards/combined_reward/std": 2.0206917849454014, "step": 29430 }, { "clip_ratio/high_max": 0.0023260214298109834, "clip_ratio/high_mean": 0.0012469184133806267, "clip_ratio/low_mean": 0.0009060798921685394, "clip_ratio/low_min": 0.00022626640614665424, "clip_ratio/region_mean": 0.0021529983130070224, "completions/clipped_ratio": 0.0, "completions/max_length": 487.6, "completions/max_terminated_length": 487.6, "completions/mean_length": 201.95911865234376, "completions/mean_terminated_length": 201.95911865234376, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.028783124561111133, "epoch": 18.029996939087848, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3015592990.0, "reward": 7.906933164596557, "reward_std": 0.33101717829704286, "rewards/combined_reward/mean": 7.906932735443116, "rewards/combined_reward/std": 2.490158128738403, "step": 29460 }, { "clip_ratio/high_max": 0.0025614059013605583, "clip_ratio/high_mean": 0.0014203576052144247, "clip_ratio/low_mean": 0.0011638637456295934, "clip_ratio/low_min": 0.00033495463139843196, "clip_ratio/region_mean": 0.0025842213954698915, "completions/clipped_ratio": 0.0, "completions/max_length": 654.1, "completions/max_terminated_length": 654.1, "completions/mean_length": 208.45599517822265, "completions/mean_terminated_length": 208.45599517822265, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.031402451389779644, "epoch": 18.048362411998777, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 3018693621.0, "reward": 8.085312509536744, "reward_std": 0.30999236553907394, "rewards/combined_reward/mean": 8.08531231880188, "rewards/combined_reward/std": 2.305150842666626, "step": 29490 }, { "clip_ratio/high_max": 0.002351569602130136, "clip_ratio/high_mean": 0.0014503077434104247, "clip_ratio/low_mean": 0.0010008456863943138, "clip_ratio/low_min": 0.00030913738834594066, "clip_ratio/region_mean": 0.0024511534274400523, "completions/clipped_ratio": 0.0, "completions/max_length": 666.7, "completions/max_terminated_length": 666.7, "completions/mean_length": 200.3885467529297, "completions/mean_terminated_length": 200.3885467529297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.031733775480339926, "epoch": 18.066727884909703, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3021763273.0, "reward": 8.031054925918578, "reward_std": 0.330524879693985, "rewards/combined_reward/mean": 8.031054735183716, "rewards/combined_reward/std": 2.2349377274513245, "step": 29520 }, { "clip_ratio/high_max": 0.00263304335870392, "clip_ratio/high_mean": 0.001450040777369092, "clip_ratio/low_mean": 0.001084788363489982, "clip_ratio/low_min": 0.0003968955070983308, "clip_ratio/region_mean": 0.0025348291548046593, "completions/clipped_ratio": 0.0, "completions/max_length": 538.5, "completions/max_terminated_length": 538.5, "completions/mean_length": 202.1671936035156, "completions/mean_terminated_length": 202.1671936035156, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.03072104329864184, "epoch": 18.085093357820632, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.003143310546875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3024839755.0, "reward": 8.076953363418578, "reward_std": 0.2894409991800785, "rewards/combined_reward/mean": 8.076953268051147, "rewards/combined_reward/std": 2.345699667930603, "step": 29550 }, { "clip_ratio/high_max": 0.0019137702882289886, "clip_ratio/high_mean": 0.001136395767510597, "clip_ratio/low_mean": 0.0008527723911659753, "clip_ratio/low_min": 0.0002959510345438806, "clip_ratio/region_mean": 0.0019891681544928967, "completions/clipped_ratio": 0.0, "completions/max_length": 532.4, "completions/max_terminated_length": 532.4, "completions/mean_length": 197.97760925292968, "completions/mean_terminated_length": 197.97760925292968, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.024788778267490368, "epoch": 18.103458830731558, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3027900149.0, "reward": 8.12485728263855, "reward_std": 0.36074853837490084, "rewards/combined_reward/mean": 8.124857091903687, "rewards/combined_reward/std": 2.2042316198349, "step": 29580 }, { "clip_ratio/high_max": 0.0022766019760941465, "clip_ratio/high_mean": 0.0012429326589578219, "clip_ratio/low_mean": 0.0009807745559858934, "clip_ratio/low_min": 0.00026039459771709517, "clip_ratio/region_mean": 0.002223707218945492, "completions/clipped_ratio": 0.0, "completions/max_length": 597.4, "completions/max_terminated_length": 597.4, "completions/mean_length": 199.0682342529297, "completions/mean_terminated_length": 199.0682342529297, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02721111143473536, "epoch": 18.121824303642484, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3030964731.0, "reward": 8.181736660003661, "reward_std": 0.309861832857132, "rewards/combined_reward/mean": 8.181736326217651, "rewards/combined_reward/std": 2.1840378642082214, "step": 29610 }, { "clip_ratio/high_max": 0.0021285014699969906, "clip_ratio/high_mean": 0.0012884795190378402, "clip_ratio/low_mean": 0.0008945070296855798, "clip_ratio/low_min": 0.0003201990468369331, "clip_ratio/region_mean": 0.0021829865319887174, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 202.15833892822266, "completions/mean_terminated_length": 202.15833892822266, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.027626965148374438, "epoch": 18.140189776553413, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3034041179.0, "reward": 8.247066307067872, "reward_std": 0.34663464576005937, "rewards/combined_reward/mean": 8.24706597328186, "rewards/combined_reward/std": 2.047052729129791, "step": 29640 }, { "clip_ratio/high_max": 0.0035632739289818954, "clip_ratio/high_mean": 0.0020057023810901835, "clip_ratio/low_mean": 0.0009784323657489344, "clip_ratio/low_min": 0.0002545170835219324, "clip_ratio/region_mean": 0.0029841347379260695, "completions/clipped_ratio": 0.0, "completions/max_length": 700.5, "completions/max_terminated_length": 700.5, "completions/mean_length": 201.6656295776367, "completions/mean_terminated_length": 201.6656295776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03662278608729442, "epoch": 18.15855524946434, "frac_reward_zero_std": 0.22083333618938922, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3037115735.0, "reward": 8.023879528045654, "reward_std": 0.36338321790099143, "rewards/combined_reward/mean": 8.023879051208496, "rewards/combined_reward/std": 2.1813615441322325, "step": 29670 }, { "clip_ratio/high_max": 0.002657680553481138, "clip_ratio/high_mean": 0.001495708951309401, "clip_ratio/low_mean": 0.000917419497394197, "clip_ratio/low_min": 0.00020077967662170219, "clip_ratio/region_mean": 0.002413128447369672, "completions/clipped_ratio": 0.0, "completions/max_length": 577.8, "completions/max_terminated_length": 577.8, "completions/mean_length": 188.65911865234375, "completions/mean_terminated_length": 188.65911865234375, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.031052446722363434, "epoch": 18.176920722375268, "frac_reward_zero_std": 0.2083333373069763, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3040140346.0, "reward": 7.7446599960327145, "reward_std": 0.2533677212893963, "rewards/combined_reward/mean": 7.74465970993042, "rewards/combined_reward/std": 2.6927117466926576, "step": 29700 }, { "clip_ratio/high_max": 0.002485601424511212, "clip_ratio/high_mean": 0.0013809582378598861, "clip_ratio/low_mean": 0.0008731133360318685, "clip_ratio/low_min": 0.0003070539103646297, "clip_ratio/region_mean": 0.0022540715707388397, "completions/clipped_ratio": 0.0, "completions/max_length": 644.5, "completions/max_terminated_length": 644.5, "completions/mean_length": 195.79271240234374, "completions/mean_terminated_length": 195.79271240234374, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.02507914728485048, "epoch": 18.195286195286194, "frac_reward_zero_std": 0.22500001192092894, "grad_norm": 0.0130615234375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3043192350.0, "reward": 7.968746852874756, "reward_std": 0.30927333384752276, "rewards/combined_reward/mean": 7.9687464237213135, "rewards/combined_reward/std": 2.504189610481262, "step": 29730 }, { "clip_ratio/high_max": 0.0024485915530628216, "clip_ratio/high_mean": 0.0014306323196554634, "clip_ratio/low_mean": 0.0010564544891773645, "clip_ratio/low_min": 0.00025050451537633, "clip_ratio/region_mean": 0.002487086820959424, "completions/clipped_ratio": 0.0, "completions/max_length": 580.1, "completions/max_terminated_length": 580.1, "completions/mean_length": 201.9976608276367, "completions/mean_terminated_length": 201.9976608276367, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.02975742716807872, "epoch": 18.213651668197123, "frac_reward_zero_std": 0.1375000048428774, "grad_norm": 0.00732421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3046268181.0, "reward": 8.270556306838989, "reward_std": 0.3331509500741959, "rewards/combined_reward/mean": 8.270556020736695, "rewards/combined_reward/std": 2.0910520911216737, "step": 29760 }, { "clip_ratio/high_max": 0.002885044735739939, "clip_ratio/high_mean": 0.0016117564320059805, "clip_ratio/low_mean": 0.0010071347326932785, "clip_ratio/low_min": 0.00022400465725998706, "clip_ratio/region_mean": 0.002618891161788876, "completions/clipped_ratio": 0.0, "completions/max_length": 580.3, "completions/max_terminated_length": 580.3, "completions/mean_length": 198.12656860351564, "completions/mean_terminated_length": 198.12656860351564, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.030007119088744125, "epoch": 18.23201714110805, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3049329147.0, "reward": 7.924990892410278, "reward_std": 0.32089762687683104, "rewards/combined_reward/mean": 7.924990510940551, "rewards/combined_reward/std": 2.355079698562622, "step": 29790 }, { "clip_ratio/high_max": 0.002083828121734162, "clip_ratio/high_mean": 0.0012981189986021491, "clip_ratio/low_mean": 0.0010601283993310062, "clip_ratio/low_min": 0.00036815243195936396, "clip_ratio/region_mean": 0.00235824740860456, "completions/clipped_ratio": 0.0, "completions/max_length": 611.6, "completions/max_terminated_length": 611.6, "completions/mean_length": 195.59948272705077, "completions/mean_terminated_length": 195.59948272705077, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.02850191555917263, "epoch": 18.250382614018978, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3052380409.0, "reward": 8.12638030052185, "reward_std": 0.2651627689599991, "rewards/combined_reward/mean": 8.126380062103271, "rewards/combined_reward/std": 2.2265942335128783, "step": 29820 }, { "clip_ratio/high_max": 0.0021822506498817046, "clip_ratio/high_mean": 0.001294178683019709, "clip_ratio/low_mean": 0.0009667522807527954, "clip_ratio/low_min": 0.00032775736593369705, "clip_ratio/region_mean": 0.002260930965227696, "completions/clipped_ratio": 0.0, "completions/max_length": 533.1, "completions/max_terminated_length": 533.1, "completions/mean_length": 203.36693267822267, "completions/mean_terminated_length": 203.36693267822267, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.02937625180929899, "epoch": 18.268748086929904, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3055461498.0, "reward": 8.355346441268921, "reward_std": 0.3113093852996826, "rewards/combined_reward/mean": 8.35534601211548, "rewards/combined_reward/std": 1.8436383366584779, "step": 29850 }, { "clip_ratio/high_max": 0.002281536959829585, "clip_ratio/high_mean": 0.0012172800123759467, "clip_ratio/low_mean": 0.0008068223983476249, "clip_ratio/low_min": 0.0001748737777234055, "clip_ratio/region_mean": 0.0020241023987182414, "completions/clipped_ratio": 0.0, "completions/max_length": 716.3, "completions/max_terminated_length": 716.3, "completions/mean_length": 202.11302337646484, "completions/mean_terminated_length": 202.11302337646484, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.02627410284864406, "epoch": 18.287113559840833, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.023193359375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3058537772.0, "reward": 8.070949077606201, "reward_std": 0.34556756168603897, "rewards/combined_reward/mean": 8.07094841003418, "rewards/combined_reward/std": 2.1624524354934693, "step": 29880 }, { "clip_ratio/high_max": 0.002450444912149881, "clip_ratio/high_mean": 0.0013876534584899977, "clip_ratio/low_mean": 0.0007783182443138988, "clip_ratio/low_min": 0.00017309641140551927, "clip_ratio/region_mean": 0.0021659717264507585, "completions/clipped_ratio": 0.0, "completions/max_length": 620.1, "completions/max_terminated_length": 620.1, "completions/mean_length": 187.7593795776367, "completions/mean_terminated_length": 187.7593795776367, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027342300194626055, "epoch": 18.30547903275176, "frac_reward_zero_std": 0.20833333618938923, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3061558928.0, "reward": 8.395206403732299, "reward_std": 0.33833037614822387, "rewards/combined_reward/mean": 8.395206260681153, "rewards/combined_reward/std": 1.9060704588890076, "step": 29910 }, { "clip_ratio/high_max": 0.002477239979392228, "clip_ratio/high_mean": 0.0014032894076080992, "clip_ratio/low_mean": 0.0009374613229131986, "clip_ratio/low_min": 0.00030935706478582385, "clip_ratio/region_mean": 0.002340750711058111, "completions/clipped_ratio": 0.0, "completions/max_length": 639.6, "completions/max_terminated_length": 639.6, "completions/mean_length": 201.97344055175782, "completions/mean_terminated_length": 201.97344055175782, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.0296885727904737, "epoch": 18.323844505662688, "frac_reward_zero_std": 0.1666666679084301, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3064634666.0, "reward": 7.798921775817871, "reward_std": 0.29365152716636655, "rewards/combined_reward/mean": 7.7989216327667235, "rewards/combined_reward/std": 2.5961947083473205, "step": 29940 }, { "clip_ratio/high_max": 0.002238131133102191, "clip_ratio/high_mean": 0.0013462847540116247, "clip_ratio/low_mean": 0.000885571890224431, "clip_ratio/low_min": 0.00023180848390135604, "clip_ratio/region_mean": 0.002231856646540109, "completions/clipped_ratio": 0.0, "completions/max_length": 522.1, "completions/max_terminated_length": 522.1, "completions/mean_length": 201.01406860351562, "completions/mean_terminated_length": 201.01406860351562, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.028518752831344803, "epoch": 18.342209978573614, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3067706720.0, "reward": 8.07209815979004, "reward_std": 0.2941159397363663, "rewards/combined_reward/mean": 8.072097873687744, "rewards/combined_reward/std": 2.3842503309249876, "step": 29970 }, { "clip_ratio/high_max": 0.0025706381275085732, "clip_ratio/high_mean": 0.0013068411019048654, "clip_ratio/low_mean": 0.0009173774121639629, "clip_ratio/low_min": 0.00025748601862384625, "clip_ratio/region_mean": 0.002224218521344786, "completions/clipped_ratio": 0.0, "completions/max_length": 646.7, "completions/max_terminated_length": 646.7, "completions/mean_length": 205.30990142822264, "completions/mean_terminated_length": 205.30990142822264, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.02904864785571893, "epoch": 18.360575451484543, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.00604248046875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3070795270.0, "reward": 8.173490715026855, "reward_std": 0.33624189496040346, "rewards/combined_reward/mean": 8.17349009513855, "rewards/combined_reward/std": 2.087234926223755, "step": 30000 }, { "clip_ratio/high_max": 0.002448122328254006, "clip_ratio/high_mean": 0.0014294412707386074, "clip_ratio/low_mean": 0.001049226917408911, "clip_ratio/low_min": 0.00030254354375453354, "clip_ratio/region_mean": 0.002478668177112316, "completions/clipped_ratio": 0.0, "completions/max_length": 530.8, "completions/max_terminated_length": 530.8, "completions/mean_length": 200.0471420288086, "completions/mean_terminated_length": 200.0471420288086, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028953596437349915, "epoch": 18.37894092439547, "frac_reward_zero_std": 0.18333333656191825, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3073863611.0, "reward": 8.249130725860596, "reward_std": 0.3309384673833847, "rewards/combined_reward/mean": 8.249130487442017, "rewards/combined_reward/std": 1.9726526498794557, "step": 30030 }, { "clip_ratio/high_max": 0.0026181445665618714, "clip_ratio/high_mean": 0.0015106057423205737, "clip_ratio/low_mean": 0.0008646424320128669, "clip_ratio/low_min": 0.00019634041527751833, "clip_ratio/region_mean": 0.0023752481623281105, "completions/clipped_ratio": 0.0, "completions/max_length": 644.1, "completions/max_terminated_length": 644.1, "completions/mean_length": 194.80833587646484, "completions/mean_terminated_length": 194.80833587646484, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027754110842943193, "epoch": 18.397306397306398, "frac_reward_zero_std": 0.25000000596046446, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3076911835.0, "reward": 7.917864274978638, "reward_std": 0.3684894278645515, "rewards/combined_reward/mean": 7.917864036560059, "rewards/combined_reward/std": 2.3858951687812806, "step": 30060 }, { "clip_ratio/high_max": 0.002307043188193347, "clip_ratio/high_mean": 0.0012840699280786793, "clip_ratio/low_mean": 0.0010245669159000196, "clip_ratio/low_min": 0.0002878689692200472, "clip_ratio/region_mean": 0.00230863684215971, "completions/clipped_ratio": 0.0, "completions/max_length": 589.7, "completions/max_terminated_length": 589.7, "completions/mean_length": 208.50130767822264, "completions/mean_terminated_length": 208.50130767822264, "completions/min_length": 39.2, "completions/min_terminated_length": 39.2, "entropy": 0.029095824419831235, "epoch": 18.415671870217324, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3080012640.0, "reward": 8.030007028579712, "reward_std": 0.329553085565567, "rewards/combined_reward/mean": 8.030006551742554, "rewards/combined_reward/std": 2.2812400341033934, "step": 30090 }, { "clip_ratio/high_max": 0.0022645239955939663, "clip_ratio/high_mean": 0.0013052634236373705, "clip_ratio/low_mean": 0.0008343844769115094, "clip_ratio/low_min": 0.00013856654744207237, "clip_ratio/region_mean": 0.0021396479147369972, "completions/clipped_ratio": 0.0, "completions/max_length": 554.3, "completions/max_terminated_length": 554.3, "completions/mean_length": 191.63255615234374, "completions/mean_terminated_length": 191.63255615234374, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.028236699492360153, "epoch": 18.434037343128253, "frac_reward_zero_std": 0.1791666679084301, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3083048669.0, "reward": 8.217516994476318, "reward_std": 0.320695810765028, "rewards/combined_reward/mean": 8.217516660690308, "rewards/combined_reward/std": 2.2095918893814086, "step": 30120 }, { "clip_ratio/high_max": 0.0021832275388684744, "clip_ratio/high_mean": 0.0011798717834608397, "clip_ratio/low_mean": 0.0008197999455054135, "clip_ratio/low_min": 0.00016834683071162242, "clip_ratio/region_mean": 0.001999671741092849, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 193.25234985351562, "completions/mean_terminated_length": 193.25234985351562, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02662409784582754, "epoch": 18.45240281603918, "frac_reward_zero_std": 0.19166666977107524, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3086090918.0, "reward": 8.124476909637451, "reward_std": 0.26452235132455826, "rewards/combined_reward/mean": 8.124476766586303, "rewards/combined_reward/std": 2.298394703865051, "step": 30150 }, { "clip_ratio/high_max": 0.0021727914609073196, "clip_ratio/high_mean": 0.0012056514173309552, "clip_ratio/low_mean": 0.0009546123657249457, "clip_ratio/low_min": 0.0003040477720787749, "clip_ratio/region_mean": 0.002160263782328305, "completions/clipped_ratio": 0.0, "completions/max_length": 617.7, "completions/max_terminated_length": 617.7, "completions/mean_length": 203.05859832763673, "completions/mean_terminated_length": 203.05859832763673, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.027171764709055423, "epoch": 18.470768288950108, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3089170823.0, "reward": 7.968183469772339, "reward_std": 0.3504512503743172, "rewards/combined_reward/mean": 7.968182849884033, "rewards/combined_reward/std": 2.3707836151123045, "step": 30180 }, { "clip_ratio/high_max": 0.0023228668675680333, "clip_ratio/high_mean": 0.0013175798743759516, "clip_ratio/low_mean": 0.0009081923653866397, "clip_ratio/low_min": 0.00019168087965226732, "clip_ratio/region_mean": 0.0022257722579524853, "completions/clipped_ratio": 0.0, "completions/max_length": 576.5, "completions/max_terminated_length": 576.5, "completions/mean_length": 200.80521240234376, "completions/mean_terminated_length": 200.80521240234376, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028027662083817025, "epoch": 18.489133761861034, "frac_reward_zero_std": 0.19166667200624943, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3092242075.0, "reward": 8.097474622726441, "reward_std": 0.3563885286450386, "rewards/combined_reward/mean": 8.097474241256714, "rewards/combined_reward/std": 2.173988378047943, "step": 30210 }, { "clip_ratio/high_max": 0.002032856620886984, "clip_ratio/high_mean": 0.0010868671802503134, "clip_ratio/low_mean": 0.0009461107600751954, "clip_ratio/low_min": 0.0003204162841332921, "clip_ratio/region_mean": 0.0020329779377789236, "completions/clipped_ratio": 0.0, "completions/max_length": 499.2, "completions/max_terminated_length": 499.2, "completions/mean_length": 190.21875457763673, "completions/mean_terminated_length": 190.21875457763673, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.024266026075929405, "epoch": 18.507499234771963, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3095272675.0, "reward": 8.0422993183136, "reward_std": 0.3276186749339104, "rewards/combined_reward/mean": 8.042299032211304, "rewards/combined_reward/std": 2.4692177534103394, "step": 30240 }, { "clip_ratio/high_max": 0.0019593024839802335, "clip_ratio/high_mean": 0.0011756251141681181, "clip_ratio/low_mean": 0.0008626814927993109, "clip_ratio/low_min": 0.00027767446251042807, "clip_ratio/region_mean": 0.002038306603693248, "completions/clipped_ratio": 0.0, "completions/max_length": 649.6, "completions/max_terminated_length": 649.6, "completions/mean_length": 209.84375610351563, "completions/mean_terminated_length": 209.84375610351563, "completions/min_length": 46.5, "completions/min_terminated_length": 46.5, "entropy": 0.025929249854137502, "epoch": 18.52586470768289, "frac_reward_zero_std": 0.1916666693985462, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3098378635.0, "reward": 8.087021493911744, "reward_std": 0.3239428400993347, "rewards/combined_reward/mean": 8.0870210647583, "rewards/combined_reward/std": 2.13486967086792, "step": 30270 }, { "clip_ratio/high_max": 0.002401872964401264, "clip_ratio/high_mean": 0.001380992061604047, "clip_ratio/low_mean": 0.0010673526187019889, "clip_ratio/low_min": 0.0003145937303391596, "clip_ratio/region_mean": 0.002448344683701483, "completions/clipped_ratio": 0.0, "completions/max_length": 587.5, "completions/max_terminated_length": 587.5, "completions/mean_length": 197.43151397705077, "completions/mean_terminated_length": 197.43151397705077, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.029737200898428757, "epoch": 18.544230180593818, "frac_reward_zero_std": 0.16666667386889458, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 3101436932.0, "reward": 8.041226816177367, "reward_std": 0.2916151225566864, "rewards/combined_reward/mean": 8.041226387023926, "rewards/combined_reward/std": 2.303371179103851, "step": 30300 }, { "clip_ratio/high_max": 0.002967031436855905, "clip_ratio/high_mean": 0.0015718455246921319, "clip_ratio/low_mean": 0.0011302207899764956, "clip_ratio/low_min": 0.00025841276971429276, "clip_ratio/region_mean": 0.002702066316851415, "completions/clipped_ratio": 0.0, "completions/max_length": 551.4, "completions/max_terminated_length": 551.4, "completions/mean_length": 193.0705780029297, "completions/mean_terminated_length": 193.0705780029297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03184322949188451, "epoch": 18.562595653504744, "frac_reward_zero_std": 0.20000000298023224, "grad_norm": 0.005859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3104478483.0, "reward": 8.128045034408569, "reward_std": 0.32053606808185575, "rewards/combined_reward/mean": 8.12804479598999, "rewards/combined_reward/std": 2.2346768975257874, "step": 30330 }, { "clip_ratio/high_max": 0.002216250286437571, "clip_ratio/high_mean": 0.0013288060618530531, "clip_ratio/low_mean": 0.0009387930173640294, "clip_ratio/low_min": 0.00029192667934694325, "clip_ratio/region_mean": 0.002267599061936683, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 186.5393295288086, "completions/mean_terminated_length": 186.5393295288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027540502557531, "epoch": 18.580961126415673, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3107494954.0, "reward": 7.815557622909546, "reward_std": 0.2842141851782799, "rewards/combined_reward/mean": 7.815557479858398, "rewards/combined_reward/std": 2.6623742818832397, "step": 30360 }, { "clip_ratio/high_max": 0.0026980187108468574, "clip_ratio/high_mean": 0.0014548687505642496, "clip_ratio/low_mean": 0.0009601853562344331, "clip_ratio/low_min": 0.0002860217951820232, "clip_ratio/region_mean": 0.0024150540877599267, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 200.6369842529297, "completions/mean_terminated_length": 200.6369842529297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03080927177021901, "epoch": 18.5993265993266, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3110565560.0, "reward": 8.0243426322937, "reward_std": 0.32449963688850403, "rewards/combined_reward/mean": 8.0243426322937, "rewards/combined_reward/std": 2.303668713569641, "step": 30390 }, { "clip_ratio/high_max": 0.0027010528661776333, "clip_ratio/high_mean": 0.001478018999963145, "clip_ratio/low_mean": 0.0009869400050471692, "clip_ratio/low_min": 0.0002687790777902895, "clip_ratio/region_mean": 0.002464959026838187, "completions/clipped_ratio": 0.0, "completions/max_length": 570.8, "completions/max_terminated_length": 570.8, "completions/mean_length": 193.46458892822267, "completions/mean_terminated_length": 193.46458892822267, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.03157229009084404, "epoch": 18.617692072237528, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3113608624.0, "reward": 8.04532732963562, "reward_std": 0.40959085524082184, "rewards/combined_reward/mean": 8.045327377319335, "rewards/combined_reward/std": 2.0370550036430357, "step": 30420 }, { "clip_ratio/high_max": 0.0026408818831744913, "clip_ratio/high_mean": 0.001526829845533939, "clip_ratio/low_mean": 0.001034827163554534, "clip_ratio/low_min": 0.00030595910914901954, "clip_ratio/region_mean": 0.002561656984228951, "completions/clipped_ratio": 0.0, "completions/max_length": 554.2, "completions/max_terminated_length": 554.2, "completions/mean_length": 198.3041748046875, "completions/mean_terminated_length": 198.3041748046875, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.03247463825779657, "epoch": 18.636057545148454, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01068115234375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3116670272.0, "reward": 8.09754467010498, "reward_std": 0.3690034806728363, "rewards/combined_reward/mean": 8.09754433631897, "rewards/combined_reward/std": 2.261354458332062, "step": 30450 }, { "clip_ratio/high_max": 0.002386203509134551, "clip_ratio/high_mean": 0.0013789148585298487, "clip_ratio/low_mean": 0.0011907170906852115, "clip_ratio/low_min": 0.0003793097287901522, "clip_ratio/region_mean": 0.002569631932419725, "completions/clipped_ratio": 0.0, "completions/max_length": 549.7, "completions/max_terminated_length": 549.7, "completions/mean_length": 206.2846420288086, "completions/mean_terminated_length": 206.2846420288086, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03188774705243607, "epoch": 18.654423018059383, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3119762565.0, "reward": 7.977036714553833, "reward_std": 0.2896813005208969, "rewards/combined_reward/mean": 7.9770365238189695, "rewards/combined_reward/std": 2.337407410144806, "step": 30480 }, { "clip_ratio/high_max": 0.0029078856993389004, "clip_ratio/high_mean": 0.0014776306389345942, "clip_ratio/low_mean": 0.0008309112632559845, "clip_ratio/low_min": 0.00017243697754262637, "clip_ratio/region_mean": 0.0023085419263225047, "completions/clipped_ratio": 0.0, "completions/max_length": 629.8, "completions/max_terminated_length": 629.8, "completions/mean_length": 197.41484832763672, "completions/mean_terminated_length": 197.41484832763672, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.032762608909979465, "epoch": 18.67278849097031, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3122820798.0, "reward": 7.99950795173645, "reward_std": 0.3704220801591873, "rewards/combined_reward/mean": 7.999507570266724, "rewards/combined_reward/std": 2.1962515711784363, "step": 30510 }, { "clip_ratio/high_max": 0.002659137751228021, "clip_ratio/high_mean": 0.0014754719944903628, "clip_ratio/low_mean": 0.0009559993173752446, "clip_ratio/low_min": 0.000200948186829919, "clip_ratio/region_mean": 0.0024314712936757132, "completions/clipped_ratio": 0.0, "completions/max_length": 537.7, "completions/max_terminated_length": 537.7, "completions/mean_length": 186.58724365234374, "completions/mean_terminated_length": 186.58724365234374, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.029419619931528966, "epoch": 18.691153963881238, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3125837453.0, "reward": 8.054789400100708, "reward_std": 0.3161818772554398, "rewards/combined_reward/mean": 8.054789352416993, "rewards/combined_reward/std": 2.246619427204132, "step": 30540 }, { "clip_ratio/high_max": 0.00228738383060166, "clip_ratio/high_mean": 0.0012821947183207763, "clip_ratio/low_mean": 0.0010394184058289586, "clip_ratio/low_min": 0.0002713821794410857, "clip_ratio/region_mean": 0.002321613111901873, "completions/clipped_ratio": 0.0, "completions/max_length": 589.5, "completions/max_terminated_length": 589.5, "completions/mean_length": 197.84818267822266, "completions/mean_terminated_length": 197.84818267822266, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029314081316503385, "epoch": 18.709519436792164, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.006805419921875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3128897350.0, "reward": 7.994704627990723, "reward_std": 0.26283292323350904, "rewards/combined_reward/mean": 7.994704532623291, "rewards/combined_reward/std": 2.4640167474746706, "step": 30570 }, { "clip_ratio/high_max": 0.0020628307761702065, "clip_ratio/high_mean": 0.0011913396658201237, "clip_ratio/low_mean": 0.0009248771291216447, "clip_ratio/low_min": 0.00023876420649078986, "clip_ratio/region_mean": 0.0021162167909399916, "completions/clipped_ratio": 0.0, "completions/max_length": 594.2, "completions/max_terminated_length": 594.2, "completions/mean_length": 201.84193420410156, "completions/mean_terminated_length": 201.84193420410156, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.029556564955661693, "epoch": 18.727884909703093, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 3131972583.0, "reward": 8.049778366088868, "reward_std": 0.370445716381073, "rewards/combined_reward/mean": 8.049778127670288, "rewards/combined_reward/std": 2.0694509744644165, "step": 30600 }, { "clip_ratio/high_max": 0.0029152240284020083, "clip_ratio/high_mean": 0.0015076935092414108, "clip_ratio/low_mean": 0.0010644897064291095, "clip_ratio/low_min": 0.00022826256511810546, "clip_ratio/region_mean": 0.002572183232405223, "completions/clipped_ratio": 0.0, "completions/max_length": 583.6, "completions/max_terminated_length": 583.6, "completions/mean_length": 198.89844360351563, "completions/mean_terminated_length": 198.89844360351563, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.032054720415423316, "epoch": 18.74625038261402, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3135036513.0, "reward": 7.556600570678711, "reward_std": 0.3482033506035805, "rewards/combined_reward/mean": 7.556600427627563, "rewards/combined_reward/std": 2.7893607020378113, "step": 30630 }, { "clip_ratio/high_max": 0.0025783879539327853, "clip_ratio/high_mean": 0.001475012600409779, "clip_ratio/low_mean": 0.0009663823118899017, "clip_ratio/low_min": 0.0003095078594924416, "clip_ratio/region_mean": 0.002441394917453484, "completions/clipped_ratio": 0.0, "completions/max_length": 609.4, "completions/max_terminated_length": 609.4, "completions/mean_length": 197.4234420776367, "completions/mean_terminated_length": 197.4234420776367, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.03084483986410002, "epoch": 18.764615855524948, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3138094779.0, "reward": 8.235033321380616, "reward_std": 0.26172383427619933, "rewards/combined_reward/mean": 8.235033082962037, "rewards/combined_reward/std": 2.18815883398056, "step": 30660 }, { "clip_ratio/high_max": 0.0023939734732266516, "clip_ratio/high_mean": 0.001399700541999967, "clip_ratio/low_mean": 0.0011003340946141786, "clip_ratio/low_min": 0.0002890411320549902, "clip_ratio/region_mean": 0.0025000346257002095, "completions/clipped_ratio": 0.0, "completions/max_length": 621.2, "completions/max_terminated_length": 621.2, "completions/mean_length": 202.57005462646484, "completions/mean_terminated_length": 202.57005462646484, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.03456796362685661, "epoch": 18.782981328435874, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3141172808.0, "reward": 8.03542652130127, "reward_std": 0.373955112695694, "rewards/combined_reward/mean": 8.03542652130127, "rewards/combined_reward/std": 2.181510496139526, "step": 30690 }, { "clip_ratio/high_max": 0.0026577392496013393, "clip_ratio/high_mean": 0.001582485109490032, "clip_ratio/low_mean": 0.0010022026096218422, "clip_ratio/low_min": 0.0002501145216228906, "clip_ratio/region_mean": 0.002584687695101214, "completions/clipped_ratio": 0.0, "completions/max_length": 582.1, "completions/max_terminated_length": 582.1, "completions/mean_length": 194.29036865234374, "completions/mean_terminated_length": 194.29036865234374, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03078066112163166, "epoch": 18.801346801346803, "frac_reward_zero_std": 0.18333334103226662, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3144219043.0, "reward": 7.94406042098999, "reward_std": 0.3553468734025955, "rewards/combined_reward/mean": 7.94406042098999, "rewards/combined_reward/std": 2.196254646778107, "step": 30720 }, { "clip_ratio/high_max": 0.0026744739661808126, "clip_ratio/high_mean": 0.0014694331989934047, "clip_ratio/low_mean": 0.0010294467745552538, "clip_ratio/low_min": 0.0002691683970624581, "clip_ratio/region_mean": 0.0024988799705170095, "completions/clipped_ratio": 0.0, "completions/max_length": 606.6, "completions/max_terminated_length": 606.6, "completions/mean_length": 210.06328430175782, "completions/mean_terminated_length": 210.06328430175782, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02997082967000703, "epoch": 18.81971227425773, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3147325846.0, "reward": 7.989070558547974, "reward_std": 0.3709010988473892, "rewards/combined_reward/mean": 7.989070463180542, "rewards/combined_reward/std": 2.28357971906662, "step": 30750 }, { "clip_ratio/high_max": 0.0017466487299922543, "clip_ratio/high_mean": 0.0009756247171026188, "clip_ratio/low_mean": 0.0009087521415494848, "clip_ratio/low_min": 0.00022114323413309953, "clip_ratio/region_mean": 0.0018843768661705933, "completions/clipped_ratio": 0.0, "completions/max_length": 630.9, "completions/max_terminated_length": 630.9, "completions/mean_length": 191.1148468017578, "completions/mean_terminated_length": 191.1148468017578, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.02507810842556258, "epoch": 18.838077747168658, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3150359887.0, "reward": 7.859599876403808, "reward_std": 0.35177039057016374, "rewards/combined_reward/mean": 7.859599828720093, "rewards/combined_reward/std": 2.4213118076324465, "step": 30780 }, { "clip_ratio/high_max": 0.002597729312158966, "clip_ratio/high_mean": 0.001521008195656274, "clip_ratio/low_mean": 0.000805429682865603, "clip_ratio/low_min": 0.00021044990353402682, "clip_ratio/region_mean": 0.002326437870215159, "completions/clipped_ratio": 0.0, "completions/max_length": 640.6, "completions/max_terminated_length": 640.6, "completions/mean_length": 208.71693267822266, "completions/mean_terminated_length": 208.71693267822266, "completions/min_length": 41.1, "completions/min_terminated_length": 41.1, "entropy": 0.028234385155762234, "epoch": 18.856443220079584, "frac_reward_zero_std": 0.20416667275130748, "grad_norm": 0.006500244140625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3153461520.0, "reward": 7.905003356933594, "reward_std": 0.32765848487615584, "rewards/combined_reward/mean": 7.905003023147583, "rewards/combined_reward/std": 2.348987710475922, "step": 30810 }, { "clip_ratio/high_max": 0.0024669522696058265, "clip_ratio/high_mean": 0.0014725326235444905, "clip_ratio/low_mean": 0.0009237338821549201, "clip_ratio/low_min": 0.00023206023251987062, "clip_ratio/region_mean": 0.0023962664946642082, "completions/clipped_ratio": 0.0, "completions/max_length": 594.4, "completions/max_terminated_length": 594.4, "completions/mean_length": 195.6526092529297, "completions/mean_terminated_length": 195.6526092529297, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.029286447446793318, "epoch": 18.874808692990513, "frac_reward_zero_std": 0.16666667386889458, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 3156512986.0, "reward": 7.983935260772705, "reward_std": 0.4342095524072647, "rewards/combined_reward/mean": 7.983935260772705, "rewards/combined_reward/std": 2.3956963777542115, "step": 30840 }, { "clip_ratio/high_max": 0.002325173378145943, "clip_ratio/high_mean": 0.0014168238325510175, "clip_ratio/low_mean": 0.0010530253131340336, "clip_ratio/low_min": 0.00040950896703482916, "clip_ratio/region_mean": 0.00246984914362353, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 213.90078582763672, "completions/mean_terminated_length": 213.90078582763672, "completions/min_length": 47.1, "completions/min_terminated_length": 47.1, "entropy": 0.030489294168849785, "epoch": 18.89317416590144, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3159634525.0, "reward": 8.103615760803223, "reward_std": 0.37775338888168336, "rewards/combined_reward/mean": 8.103615427017212, "rewards/combined_reward/std": 2.2573359489440916, "step": 30870 }, { "clip_ratio/high_max": 0.0024365715580643155, "clip_ratio/high_mean": 0.0014488017865611862, "clip_ratio/low_mean": 0.0011378366102386887, "clip_ratio/low_min": 0.0003560901765013114, "clip_ratio/region_mean": 0.0025866383885537894, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 214.8195373535156, "completions/mean_terminated_length": 214.8195373535156, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.029162115409659843, "epoch": 18.911539638812368, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3162759592.0, "reward": 7.959027481079102, "reward_std": 0.40269329994916914, "rewards/combined_reward/mean": 7.959027099609375, "rewards/combined_reward/std": 2.3050203561782836, "step": 30900 }, { "clip_ratio/high_max": 0.0022059799356308456, "clip_ratio/high_mean": 0.0013375883723104683, "clip_ratio/low_mean": 0.000906227951903323, "clip_ratio/low_min": 0.00023368499120503353, "clip_ratio/region_mean": 0.0022438163299132916, "completions/clipped_ratio": 0.0, "completions/max_length": 627.8, "completions/max_terminated_length": 627.8, "completions/mean_length": 202.33568267822267, "completions/mean_terminated_length": 202.33568267822267, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.02660148701009651, "epoch": 18.929905111723293, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3165836721.0, "reward": 8.011403226852417, "reward_std": 0.37770378589630127, "rewards/combined_reward/mean": 8.011403036117553, "rewards/combined_reward/std": 2.3638768911361696, "step": 30930 }, { "clip_ratio/high_max": 0.0028948225197382273, "clip_ratio/high_mean": 0.0017417018178093712, "clip_ratio/low_mean": 0.0011242906678186652, "clip_ratio/low_min": 0.00036134278707322665, "clip_ratio/region_mean": 0.0028659924796860044, "completions/clipped_ratio": 0.0, "completions/max_length": 609.9, "completions/max_terminated_length": 609.9, "completions/mean_length": 208.1867233276367, "completions/mean_terminated_length": 208.1867233276367, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.03232750648943086, "epoch": 18.948270584634223, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.008056640625, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 3168936318.0, "reward": 7.930130958557129, "reward_std": 0.3836293339729309, "rewards/combined_reward/mean": 7.9301307678222654, "rewards/combined_reward/std": 2.374290370941162, "step": 30960 }, { "clip_ratio/high_max": 0.0024540387295322336, "clip_ratio/high_mean": 0.0014318113512369261, "clip_ratio/low_mean": 0.0010743236491786471, "clip_ratio/low_min": 0.00031796164524469834, "clip_ratio/region_mean": 0.0025061350068426693, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 204.66094207763672, "completions/mean_terminated_length": 204.66094207763672, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.030403846603197355, "epoch": 18.96663605754515, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.036376953125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3172022376.0, "reward": 8.003809356689453, "reward_std": 0.35719977766275407, "rewards/combined_reward/mean": 8.003809118270874, "rewards/combined_reward/std": 2.3057252287864687, "step": 30990 }, { "clip_ratio/high_max": 0.002196403589914553, "clip_ratio/high_mean": 0.0012570714308822062, "clip_ratio/low_mean": 0.0008560768797906349, "clip_ratio/low_min": 0.00027518642964423633, "clip_ratio/region_mean": 0.0021131483338346396, "completions/clipped_ratio": 0.0, "completions/max_length": 578.9, "completions/max_terminated_length": 578.9, "completions/mean_length": 193.15313110351562, "completions/mean_terminated_length": 193.15313110351562, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.028445639981267354, "epoch": 18.985001530456074, "frac_reward_zero_std": 0.16250000670552253, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3175064244.0, "reward": 7.959527826309204, "reward_std": 0.33560983687639234, "rewards/combined_reward/mean": 7.959527635574341, "rewards/combined_reward/std": 2.5728008151054382, "step": 31020 }, { "clip_ratio/high_max": 0.0026838363760439046, "clip_ratio/high_mean": 0.0016283955012760874, "clip_ratio/low_mean": 0.0009975034271966728, "clip_ratio/low_min": 0.0003078655846381434, "clip_ratio/region_mean": 0.0026258989326040246, "completions/clipped_ratio": 0.0, "completions/max_length": 545.4444444444445, "completions/max_terminated_length": 545.4444444444445, "completions/mean_length": 195.75955369737414, "completions/mean_terminated_length": 195.75955369737414, "completions/min_length": 37.77777777777778, "completions/min_terminated_length": 37.77777777777778, "entropy": 0.03263724560581022, "epoch": 19.00306091215182, "frac_reward_zero_std": 0.14814815173546472, "grad_norm": 0.055419921875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3177810933.0, "reward": 8.264380243089464, "reward_std": 0.34912238352828556, "rewards/combined_reward/mean": 8.264379872216118, "rewards/combined_reward/std": 1.938479741414388, "step": 31050 }, { "clip_ratio/high_max": 0.0022003407415468247, "clip_ratio/high_mean": 0.0012112990341847762, "clip_ratio/low_mean": 0.0007690490272580063, "clip_ratio/low_min": 0.00025636556990017804, "clip_ratio/region_mean": 0.001980348051195809, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 197.41328887939454, "completions/mean_terminated_length": 197.41328887939454, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03170141524945696, "epoch": 19.02142638506275, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3180869160.0, "reward": 7.968004846572876, "reward_std": 0.3565476626157761, "rewards/combined_reward/mean": 7.968004417419434, "rewards/combined_reward/std": 2.4158833980560304, "step": 31080 }, { "clip_ratio/high_max": 0.002158643088963193, "clip_ratio/high_mean": 0.001222238662254919, "clip_ratio/low_mean": 0.0005527261867731189, "clip_ratio/low_min": 0.00011545149730712486, "clip_ratio/region_mean": 0.0017749648284128246, "completions/clipped_ratio": 0.0, "completions/max_length": 615.1, "completions/max_terminated_length": 615.1, "completions/mean_length": 204.90573425292968, "completions/mean_terminated_length": 204.90573425292968, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.029992589517496525, "epoch": 19.039791857973675, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3183956158.0, "reward": 7.888389635086059, "reward_std": 0.38795594573020936, "rewards/combined_reward/mean": 7.88838939666748, "rewards/combined_reward/std": 2.243713140487671, "step": 31110 }, { "clip_ratio/high_max": 0.002013360426159731, "clip_ratio/high_mean": 0.0010983682149041365, "clip_ratio/low_mean": 0.0007093272231334898, "clip_ratio/low_min": 0.00021918846420400465, "clip_ratio/region_mean": 0.0018076954045682214, "completions/clipped_ratio": 0.0, "completions/max_length": 614.3, "completions/max_terminated_length": 614.3, "completions/mean_length": 192.72604522705078, "completions/mean_terminated_length": 192.72604522705078, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030949125764891505, "epoch": 19.058157330884605, "frac_reward_zero_std": 0.21666667535901069, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3186996386.0, "reward": 7.846942949295044, "reward_std": 0.3684666246175766, "rewards/combined_reward/mean": 7.846942567825318, "rewards/combined_reward/std": 2.3287840366363524, "step": 31140 }, { "clip_ratio/high_max": 0.001967907637784568, "clip_ratio/high_mean": 0.0010485244912464017, "clip_ratio/low_mean": 0.0009032830479554832, "clip_ratio/low_min": 0.000334418328323712, "clip_ratio/region_mean": 0.0019518075316833953, "completions/clipped_ratio": 0.0, "completions/max_length": 494.7, "completions/max_terminated_length": 494.7, "completions/mean_length": 201.6242218017578, "completions/mean_terminated_length": 201.6242218017578, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.02949071992188692, "epoch": 19.07652280379553, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3190070783.0, "reward": 8.311982488632202, "reward_std": 0.40220153480768206, "rewards/combined_reward/mean": 8.311982345581054, "rewards/combined_reward/std": 2.0731898307800294, "step": 31170 }, { "clip_ratio/high_max": 0.0017517867313775544, "clip_ratio/high_mean": 0.0010051093417132507, "clip_ratio/low_mean": 0.0007239947994094109, "clip_ratio/low_min": 0.00020546802819202032, "clip_ratio/region_mean": 0.0017291041425778531, "completions/clipped_ratio": 0.0, "completions/max_length": 604.8, "completions/max_terminated_length": 604.8, "completions/mean_length": 200.50443420410156, "completions/mean_terminated_length": 200.50443420410156, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02854504665204634, "epoch": 19.09488827670646, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3193140880.0, "reward": 8.011460065841675, "reward_std": 0.3264983341097832, "rewards/combined_reward/mean": 8.011459732055664, "rewards/combined_reward/std": 2.422873246669769, "step": 31200 }, { "clip_ratio/high_max": 0.0018859375384636223, "clip_ratio/high_mean": 0.0011068498166423523, "clip_ratio/low_mean": 0.0007590579768778601, "clip_ratio/low_min": 0.00016901173124400277, "clip_ratio/region_mean": 0.0018659077783619675, "completions/clipped_ratio": 0.0, "completions/max_length": 570.2, "completions/max_terminated_length": 570.2, "completions/mean_length": 193.6320343017578, "completions/mean_terminated_length": 193.6320343017578, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.027615125221200288, "epoch": 19.112641567187023, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3196184587.0, "reward": 7.9652611255645756, "reward_std": 0.3790230453014374, "rewards/combined_reward/mean": 7.965260934829712, "rewards/combined_reward/std": 2.415210545063019, "step": 31230 }, { "clip_ratio/high_max": 0.0018134528969919, "clip_ratio/high_mean": 0.0010579635595301322, "clip_ratio/low_mean": 0.0008163564905771636, "clip_ratio/low_min": 0.00016369722191787636, "clip_ratio/region_mean": 0.0018743200358585454, "completions/clipped_ratio": 0.0, "completions/max_length": 511.1, "completions/max_terminated_length": 511.1, "completions/mean_length": 198.1080764770508, "completions/mean_terminated_length": 198.1080764770508, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.033040953365465, "epoch": 19.13100704009795, "frac_reward_zero_std": 0.1625000022351742, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3199245482.0, "reward": 7.875048494338989, "reward_std": 0.37977318167686464, "rewards/combined_reward/mean": 7.875048112869263, "rewards/combined_reward/std": 2.600023663043976, "step": 31260 }, { "clip_ratio/high_max": 0.00201197595645984, "clip_ratio/high_mean": 0.0011201266363059404, "clip_ratio/low_mean": 0.0007930905526639738, "clip_ratio/low_min": 0.00024298013959196397, "clip_ratio/region_mean": 0.001913217173811669, "completions/clipped_ratio": 0.0, "completions/max_length": 586.1, "completions/max_terminated_length": 586.1, "completions/mean_length": 209.64583892822264, "completions/mean_terminated_length": 209.64583892822264, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.030887054512277246, "epoch": 19.149372513008878, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.00726318359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3202350682.0, "reward": 8.080538082122803, "reward_std": 0.46262651681900024, "rewards/combined_reward/mean": 8.080537700653077, "rewards/combined_reward/std": 2.2609591722488402, "step": 31290 }, { "clip_ratio/high_max": 0.002138269985152874, "clip_ratio/high_mean": 0.0012792705946291485, "clip_ratio/low_mean": 0.000952772391125715, "clip_ratio/low_min": 0.0002324555124990487, "clip_ratio/region_mean": 0.0022320430134035027, "completions/clipped_ratio": 0.0, "completions/max_length": 589.2, "completions/max_terminated_length": 589.2, "completions/mean_length": 196.21380462646485, "completions/mean_terminated_length": 196.21380462646485, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.03259641374461353, "epoch": 19.167737985919803, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.0299072265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3205404303.0, "reward": 7.9374157905578615, "reward_std": 0.4056299701333046, "rewards/combined_reward/mean": 7.937415409088135, "rewards/combined_reward/std": 2.3013227581977844, "step": 31320 }, { "clip_ratio/high_max": 0.0019015853399954115, "clip_ratio/high_mean": 0.0010552134522489117, "clip_ratio/low_mean": 0.000965391237938699, "clip_ratio/low_min": 0.0002799301034732101, "clip_ratio/region_mean": 0.002020604655263014, "completions/clipped_ratio": 0.0, "completions/max_length": 548.3, "completions/max_terminated_length": 548.3, "completions/mean_length": 199.2476593017578, "completions/mean_terminated_length": 199.2476593017578, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.03152815227707227, "epoch": 19.186103458830733, "frac_reward_zero_std": 0.18333333618938924, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3208469574.0, "reward": 7.862385940551758, "reward_std": 0.34331234693527224, "rewards/combined_reward/mean": 7.862385845184326, "rewards/combined_reward/std": 2.4923221707344054, "step": 31350 }, { "clip_ratio/high_max": 0.0020337904948974027, "clip_ratio/high_mean": 0.0011385935281092922, "clip_ratio/low_mean": 0.0006853745791886468, "clip_ratio/low_min": 0.00012777225298729415, "clip_ratio/region_mean": 0.0018239681094807262, "completions/clipped_ratio": 0.0, "completions/max_length": 575.1, "completions/max_terminated_length": 575.1, "completions/mean_length": 200.19922332763673, "completions/mean_terminated_length": 200.19922332763673, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03080784309034546, "epoch": 19.20446893174166, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3211538499.0, "reward": 8.012182378768921, "reward_std": 0.3765489488840103, "rewards/combined_reward/mean": 8.01218228340149, "rewards/combined_reward/std": 2.3761888265609743, "step": 31380 }, { "clip_ratio/high_max": 0.00206334913673345, "clip_ratio/high_mean": 0.0011977201810320062, "clip_ratio/low_mean": 0.000788594057909601, "clip_ratio/low_min": 0.0002257675650374343, "clip_ratio/region_mean": 0.001986314231180586, "completions/clipped_ratio": 0.0, "completions/max_length": 594.1, "completions/max_terminated_length": 594.1, "completions/mean_length": 196.26693267822264, "completions/mean_terminated_length": 196.26693267822264, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02939402712509036, "epoch": 19.222834404652588, "frac_reward_zero_std": 0.17083334028720856, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3214592324.0, "reward": 8.209515953063965, "reward_std": 0.39280837774276733, "rewards/combined_reward/mean": 8.2095157623291, "rewards/combined_reward/std": 2.0744145274162293, "step": 31410 }, { "clip_ratio/high_max": 0.002530879754340276, "clip_ratio/high_mean": 0.0014469462340154374, "clip_ratio/low_mean": 0.0009334135764220263, "clip_ratio/low_min": 0.00023308490296282495, "clip_ratio/region_mean": 0.0023803598024339104, "completions/clipped_ratio": 0.0, "completions/max_length": 553.6, "completions/max_terminated_length": 553.6, "completions/mean_length": 211.3151092529297, "completions/mean_terminated_length": 211.3151092529297, "completions/min_length": 48.3, "completions/min_terminated_length": 48.3, "entropy": 0.035448320722207424, "epoch": 19.241199877563513, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0096435546875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3217703934.0, "reward": 8.003818607330322, "reward_std": 0.30972478836774825, "rewards/combined_reward/mean": 8.003818464279174, "rewards/combined_reward/std": 2.184041237831116, "step": 31440 }, { "clip_ratio/high_max": 0.002119336408941308, "clip_ratio/high_mean": 0.001132970360777108, "clip_ratio/low_mean": 0.0007347500726609723, "clip_ratio/low_min": 0.0002269273415246668, "clip_ratio/region_mean": 0.0018677204381674529, "completions/clipped_ratio": 0.0, "completions/max_length": 666.5, "completions/max_terminated_length": 666.5, "completions/mean_length": 192.6460968017578, "completions/mean_terminated_length": 192.6460968017578, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02995434404971699, "epoch": 19.259565350474443, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3220743855.0, "reward": 7.880805969238281, "reward_std": 0.3912398457527161, "rewards/combined_reward/mean": 7.880805969238281, "rewards/combined_reward/std": 2.597106909751892, "step": 31470 }, { "clip_ratio/high_max": 0.0018344276817515493, "clip_ratio/high_mean": 0.0010619358688321275, "clip_ratio/low_mean": 0.0007698248626790398, "clip_ratio/low_min": 0.00016409669794181052, "clip_ratio/region_mean": 0.0018317607507924549, "completions/clipped_ratio": 0.0, "completions/max_length": 588.1, "completions/max_terminated_length": 588.1, "completions/mean_length": 194.9562515258789, "completions/mean_terminated_length": 194.9562515258789, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027573258041714628, "epoch": 19.27793082338537, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3223792647.0, "reward": 7.868151712417602, "reward_std": 0.3628584861755371, "rewards/combined_reward/mean": 7.868151426315308, "rewards/combined_reward/std": 2.471984899044037, "step": 31500 }, { "clip_ratio/high_max": 0.0023034643585560844, "clip_ratio/high_mean": 0.0013476066587221188, "clip_ratio/low_mean": 0.0008481152820119557, "clip_ratio/low_min": 0.00025815113743495505, "clip_ratio/region_mean": 0.002195721938430021, "completions/clipped_ratio": 0.0, "completions/max_length": 610.6, "completions/max_terminated_length": 610.6, "completions/mean_length": 216.85260925292968, "completions/mean_terminated_length": 216.85260925292968, "completions/min_length": 41.1, "completions/min_terminated_length": 41.1, "entropy": 0.033195818696791925, "epoch": 19.296296296296298, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 3226925521.0, "reward": 8.126117372512818, "reward_std": 0.35434287041425705, "rewards/combined_reward/mean": 8.12611689567566, "rewards/combined_reward/std": 2.072297763824463, "step": 31530 }, { "clip_ratio/high_max": 0.0019015312505265077, "clip_ratio/high_mean": 0.0011555840515939052, "clip_ratio/low_mean": 0.0007300643682659332, "clip_ratio/low_min": 0.00018363672691824225, "clip_ratio/region_mean": 0.0018856484088852691, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 193.72005767822264, "completions/mean_terminated_length": 193.72005767822264, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02841296271265795, "epoch": 19.314661769207223, "frac_reward_zero_std": 0.20000000447034835, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3229969566.0, "reward": 8.028454208374024, "reward_std": 0.31852284222841265, "rewards/combined_reward/mean": 8.028453826904297, "rewards/combined_reward/std": 2.2632299780845644, "step": 31560 }, { "clip_ratio/high_max": 0.0017080744573225578, "clip_ratio/high_mean": 0.000993614319061938, "clip_ratio/low_mean": 0.0007386296977832293, "clip_ratio/low_min": 0.0001988787063358662, "clip_ratio/region_mean": 0.0017322439884689326, "completions/clipped_ratio": 0.0, "completions/max_length": 533.2, "completions/max_terminated_length": 533.2, "completions/mean_length": 203.3541732788086, "completions/mean_terminated_length": 203.3541732788086, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.02856375025585294, "epoch": 19.333027242118153, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.0224609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3233050606.0, "reward": 8.065227746963501, "reward_std": 0.36682664453983305, "rewards/combined_reward/mean": 8.065227508544922, "rewards/combined_reward/std": 2.3166542649269104, "step": 31590 }, { "clip_ratio/high_max": 0.0016429308743681759, "clip_ratio/high_mean": 0.0009419175587633314, "clip_ratio/low_mean": 0.0008279604643272857, "clip_ratio/low_min": 0.00024088145388911168, "clip_ratio/region_mean": 0.0017698780405529153, "completions/clipped_ratio": 0.0, "completions/max_length": 557.6, "completions/max_terminated_length": 557.6, "completions/mean_length": 193.37630615234374, "completions/mean_terminated_length": 193.37630615234374, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.027976381468276183, "epoch": 19.35139271502908, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3236093331.0, "reward": 8.135133838653564, "reward_std": 0.2683309629559517, "rewards/combined_reward/mean": 8.135133457183837, "rewards/combined_reward/std": 2.1579300105571746, "step": 31620 }, { "clip_ratio/high_max": 0.002106739028628605, "clip_ratio/high_mean": 0.0012459400721127167, "clip_ratio/low_mean": 0.0009607477564713918, "clip_ratio/low_min": 0.00030977669545488123, "clip_ratio/region_mean": 0.002206687826158789, "completions/clipped_ratio": 0.0, "completions/max_length": 588.1, "completions/max_terminated_length": 588.1, "completions/mean_length": 199.82318115234375, "completions/mean_terminated_length": 199.82318115234375, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 0.033314845090111095, "epoch": 19.369758187940008, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3239160812.0, "reward": 8.042355012893676, "reward_std": 0.3207100391387939, "rewards/combined_reward/mean": 8.042355012893676, "rewards/combined_reward/std": 2.079729402065277, "step": 31650 }, { "clip_ratio/high_max": 0.001876551771905118, "clip_ratio/high_mean": 0.0010777950292928532, "clip_ratio/low_mean": 0.0007041617514914833, "clip_ratio/low_min": 0.00023016480336082167, "clip_ratio/region_mean": 0.0017819567886666239, "completions/clipped_ratio": 0.0, "completions/max_length": 707.5, "completions/max_terminated_length": 707.5, "completions/mean_length": 214.19427795410155, "completions/mean_terminated_length": 214.19427795410155, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02690230687148869, "epoch": 19.388123660850933, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0235595703125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3242283478.0, "reward": 8.175826740264892, "reward_std": 0.31234969943761826, "rewards/combined_reward/mean": 8.17582664489746, "rewards/combined_reward/std": 2.2956345319747924, "step": 31680 }, { "clip_ratio/high_max": 0.0018492955627152697, "clip_ratio/high_mean": 0.0009540215138258646, "clip_ratio/low_mean": 0.0008385303752826682, "clip_ratio/low_min": 0.00023517976636261057, "clip_ratio/region_mean": 0.0017925518855918198, "completions/clipped_ratio": 0.0, "completions/max_length": 571.1, "completions/max_terminated_length": 571.1, "completions/mean_length": 199.79453735351564, "completions/mean_terminated_length": 199.79453735351564, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028520246983195344, "epoch": 19.406489133761863, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.00579833984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3245350849.0, "reward": 8.296985244750976, "reward_std": 0.41315854638814925, "rewards/combined_reward/mean": 8.296985006332397, "rewards/combined_reward/std": 2.089216148853302, "step": 31710 }, { "clip_ratio/high_max": 0.001972084282412349, "clip_ratio/high_mean": 0.0010618114229146158, "clip_ratio/low_mean": 0.0008354191355465446, "clip_ratio/low_min": 0.00019410684229418014, "clip_ratio/region_mean": 0.0018972305551869795, "completions/clipped_ratio": 0.0, "completions/max_length": 615.2, "completions/max_terminated_length": 615.2, "completions/mean_length": 201.8932342529297, "completions/mean_terminated_length": 201.8932342529297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.029751868406310677, "epoch": 19.42485460667279, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3248426279.0, "reward": 8.07138671875, "reward_std": 0.33984803780913353, "rewards/combined_reward/mean": 8.071386480331421, "rewards/combined_reward/std": 2.250631403923035, "step": 31740 }, { "clip_ratio/high_max": 0.0019425373844569548, "clip_ratio/high_mean": 0.0010758423314352209, "clip_ratio/low_mean": 0.0009013296028570039, "clip_ratio/low_min": 0.0003179725557856727, "clip_ratio/region_mean": 0.0019771719555137677, "completions/clipped_ratio": 0.0, "completions/max_length": 590.1, "completions/max_terminated_length": 590.1, "completions/mean_length": 199.88359985351562, "completions/mean_terminated_length": 199.88359985351562, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.02795134865057965, "epoch": 19.443220079583718, "frac_reward_zero_std": 0.13333333507180214, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3251493992.0, "reward": 8.263382005691529, "reward_std": 0.2901636078953743, "rewards/combined_reward/mean": 8.26338176727295, "rewards/combined_reward/std": 1.9799288868904115, "step": 31770 }, { "clip_ratio/high_max": 0.0019461601419607177, "clip_ratio/high_mean": 0.0010764223060202008, "clip_ratio/low_mean": 0.0005998359720251756, "clip_ratio/low_min": 0.00012477444688556715, "clip_ratio/region_mean": 0.0016762582965990684, "completions/clipped_ratio": 0.0, "completions/max_length": 583.1, "completions/max_terminated_length": 583.1, "completions/mean_length": 196.76693420410157, "completions/mean_terminated_length": 196.76693420410157, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.026649790164083243, "epoch": 19.461585552494643, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.004119873046875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3254549737.0, "reward": 8.216329908370971, "reward_std": 0.43601962327957156, "rewards/combined_reward/mean": 8.216329622268677, "rewards/combined_reward/std": 2.035055601596832, "step": 31800 }, { "clip_ratio/high_max": 0.001961022136674728, "clip_ratio/high_mean": 0.0011214810018524682, "clip_ratio/low_mean": 0.0007498258213066341, "clip_ratio/low_min": 0.00021091013428910325, "clip_ratio/region_mean": 0.0018713068183084638, "completions/clipped_ratio": 0.0, "completions/max_length": 587.5, "completions/max_terminated_length": 587.5, "completions/mean_length": 198.20781860351562, "completions/mean_terminated_length": 198.20781860351562, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.0299223601197203, "epoch": 19.479951025405573, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3257611015.0, "reward": 8.164685916900634, "reward_std": 0.33549973741173744, "rewards/combined_reward/mean": 8.164685821533203, "rewards/combined_reward/std": 1.9908003211021423, "step": 31830 }, { "clip_ratio/high_max": 0.0017104848304976864, "clip_ratio/high_mean": 0.0010527573893341468, "clip_ratio/low_mean": 0.0006707835286457946, "clip_ratio/low_min": 0.0001851888227974996, "clip_ratio/region_mean": 0.0017235409085211964, "completions/clipped_ratio": 0.0, "completions/max_length": 533.7, "completions/max_terminated_length": 533.7, "completions/mean_length": 202.8893280029297, "completions/mean_terminated_length": 202.8893280029297, "completions/min_length": 45.7, "completions/min_terminated_length": 45.7, "entropy": 0.028846293517077962, "epoch": 19.4983164983165, "frac_reward_zero_std": 0.1333333358168602, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3260690270.0, "reward": 7.778277015686035, "reward_std": 0.3949850216507912, "rewards/combined_reward/mean": 7.77827672958374, "rewards/combined_reward/std": 2.3505916833877563, "step": 31860 }, { "clip_ratio/high_max": 0.0017971930045556897, "clip_ratio/high_mean": 0.0010176639778364915, "clip_ratio/low_mean": 0.0005919126165584506, "clip_ratio/low_min": 0.0001439313797163777, "clip_ratio/region_mean": 0.0016095765808131547, "completions/clipped_ratio": 0.0, "completions/max_length": 575.9, "completions/max_terminated_length": 575.9, "completions/mean_length": 200.91120300292968, "completions/mean_terminated_length": 200.91120300292968, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.026828807964920997, "epoch": 19.516681971227428, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.006256103515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3263761929.0, "reward": 7.913752174377441, "reward_std": 0.3329450890421867, "rewards/combined_reward/mean": 7.913752031326294, "rewards/combined_reward/std": 2.3737168312072754, "step": 31890 }, { "clip_ratio/high_max": 0.0020732689158952174, "clip_ratio/high_mean": 0.0011725316594189887, "clip_ratio/low_mean": 0.0009196217000862816, "clip_ratio/low_min": 0.0002929036752902903, "clip_ratio/region_mean": 0.0020921533393751207, "completions/clipped_ratio": 0.0, "completions/max_length": 611.3, "completions/max_terminated_length": 611.3, "completions/mean_length": 208.37656860351564, "completions/mean_terminated_length": 208.37656860351564, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.03171539013274014, "epoch": 19.535047444138353, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.005340576171875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3266862255.0, "reward": 7.639175128936768, "reward_std": 0.34496105313301084, "rewards/combined_reward/mean": 7.639175081253052, "rewards/combined_reward/std": 2.6949435472488403, "step": 31920 }, { "clip_ratio/high_max": 0.0019146982022599937, "clip_ratio/high_mean": 0.0011862984373389432, "clip_ratio/low_mean": 0.0007870053656612678, "clip_ratio/low_min": 0.00019934926822315902, "clip_ratio/region_mean": 0.0019733038027576794, "completions/clipped_ratio": 0.0, "completions/max_length": 727.1, "completions/max_terminated_length": 727.1, "completions/mean_length": 206.86875610351564, "completions/mean_terminated_length": 206.86875610351564, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.030883919354528187, "epoch": 19.55341291704928, "frac_reward_zero_std": 0.19583333991467952, "grad_norm": 0.021728515625, "learning_rate": 5e-06, "loss": 0.0029, "num_tokens": 3269956791.0, "reward": 8.237307739257812, "reward_std": 0.43126660883426665, "rewards/combined_reward/mean": 8.237307357788087, "rewards/combined_reward/std": 2.061427283287048, "step": 31950 }, { "clip_ratio/high_max": 0.00220082329227201, "clip_ratio/high_mean": 0.0012347955140285193, "clip_ratio/low_mean": 0.0007479132272540785, "clip_ratio/low_min": 0.00022115484098321757, "clip_ratio/region_mean": 0.0019827087579566676, "completions/clipped_ratio": 0.0, "completions/max_length": 533.6, "completions/max_terminated_length": 533.6, "completions/mean_length": 195.04765930175782, "completions/mean_terminated_length": 195.04765930175782, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029812683040897053, "epoch": 19.57177838996021, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3273005934.0, "reward": 7.91985297203064, "reward_std": 0.3494377747178078, "rewards/combined_reward/mean": 7.919852781295776, "rewards/combined_reward/std": 2.3569815158843994, "step": 31980 }, { "clip_ratio/high_max": 0.0019149063276321006, "clip_ratio/high_mean": 0.0010761134203979357, "clip_ratio/low_mean": 0.0008629189237884323, "clip_ratio/low_min": 0.00025658263354368197, "clip_ratio/region_mean": 0.0019390323490370066, "completions/clipped_ratio": 0.0, "completions/max_length": 558.7, "completions/max_terminated_length": 558.7, "completions/mean_length": 195.41120147705078, "completions/mean_terminated_length": 195.41120147705078, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.030084933418159685, "epoch": 19.590143862871134, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3276056473.0, "reward": 8.02201738357544, "reward_std": 0.3766194939613342, "rewards/combined_reward/mean": 8.022017335891723, "rewards/combined_reward/std": 2.273285222053528, "step": 32010 }, { "clip_ratio/high_max": 0.0016203795748879202, "clip_ratio/high_mean": 0.0008917065872083186, "clip_ratio/low_mean": 0.0009097339750345176, "clip_ratio/low_min": 0.0003290563428890891, "clip_ratio/region_mean": 0.001801440565517017, "completions/clipped_ratio": 0.0, "completions/max_length": 560.1, "completions/max_terminated_length": 560.1, "completions/mean_length": 189.8062530517578, "completions/mean_terminated_length": 189.8062530517578, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02824279269358764, "epoch": 19.608509335782063, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 3279085489.0, "reward": 8.197990798950196, "reward_std": 0.3538346067070961, "rewards/combined_reward/mean": 8.197990703582764, "rewards/combined_reward/std": 2.0427629113197328, "step": 32040 }, { "clip_ratio/high_max": 0.0018525660333883328, "clip_ratio/high_mean": 0.0010147566124942386, "clip_ratio/low_mean": 0.0007612266120607577, "clip_ratio/low_min": 0.00023299686145037413, "clip_ratio/region_mean": 0.0017759832306182943, "completions/clipped_ratio": 0.0, "completions/max_length": 594.9, "completions/max_terminated_length": 594.9, "completions/mean_length": 195.50000610351563, "completions/mean_terminated_length": 195.50000610351563, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029705777267615, "epoch": 19.62687480869299, "frac_reward_zero_std": 0.21666667014360427, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3282136369.0, "reward": 8.048191213607788, "reward_std": 0.3273344412446022, "rewards/combined_reward/mean": 8.048190784454345, "rewards/combined_reward/std": 2.509297513961792, "step": 32070 }, { "clip_ratio/high_max": 0.0018162864245823585, "clip_ratio/high_mean": 0.0010892162268040313, "clip_ratio/low_mean": 0.0006873252967731484, "clip_ratio/low_min": 0.00018983151094289496, "clip_ratio/region_mean": 0.0017765415133908391, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 190.3854232788086, "completions/mean_terminated_length": 190.3854232788086, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.028364223466875654, "epoch": 19.645240281603918, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0042724609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3285167609.0, "reward": 8.068116521835327, "reward_std": 0.29191186726093293, "rewards/combined_reward/mean": 8.0681161403656, "rewards/combined_reward/std": 2.177958881855011, "step": 32100 }, { "clip_ratio/high_max": 0.001693786987986338, "clip_ratio/high_mean": 0.0009704230580003544, "clip_ratio/low_mean": 0.0009712941026615833, "clip_ratio/low_min": 0.0003237922819001445, "clip_ratio/region_mean": 0.0019417171463525543, "completions/clipped_ratio": 0.0, "completions/max_length": 531.9, "completions/max_terminated_length": 531.9, "completions/mean_length": 207.11953582763672, "completions/mean_terminated_length": 207.11953582763672, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.028941119058678546, "epoch": 19.663605754514844, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.024658203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3288263108.0, "reward": 7.876861953735352, "reward_std": 0.3799701899290085, "rewards/combined_reward/mean": 7.876861619949341, "rewards/combined_reward/std": 2.444371592998505, "step": 32130 }, { "clip_ratio/high_max": 0.0020944480696925894, "clip_ratio/high_mean": 0.0010977870608864274, "clip_ratio/low_mean": 0.0007921329416907004, "clip_ratio/low_min": 0.00022178566772102689, "clip_ratio/region_mean": 0.0018899199965138299, "completions/clipped_ratio": 0.0, "completions/max_length": 524.9, "completions/max_terminated_length": 524.9, "completions/mean_length": 194.40912017822265, "completions/mean_terminated_length": 194.40912017822265, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.028048008548406264, "epoch": 19.681971227425773, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3291309799.0, "reward": 7.894447898864746, "reward_std": 0.2960650593042374, "rewards/combined_reward/mean": 7.894447660446167, "rewards/combined_reward/std": 2.238171875476837, "step": 32160 }, { "clip_ratio/high_max": 0.001896991108272535, "clip_ratio/high_mean": 0.0010853830351455448, "clip_ratio/low_mean": 0.000866006681098952, "clip_ratio/low_min": 0.00024331814626445217, "clip_ratio/region_mean": 0.00195138971224272, "completions/clipped_ratio": 0.0, "completions/max_length": 574.5, "completions/max_terminated_length": 574.5, "completions/mean_length": 184.72995300292968, "completions/mean_terminated_length": 184.72995300292968, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030560849715645114, "epoch": 19.7003367003367, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3294319322.0, "reward": 7.574043035507202, "reward_std": 0.3865486979484558, "rewards/combined_reward/mean": 7.5740429878234865, "rewards/combined_reward/std": 2.6423280000686646, "step": 32190 }, { "clip_ratio/high_max": 0.0015638390044235468, "clip_ratio/high_mean": 0.0008788866147369845, "clip_ratio/low_mean": 0.0007700593773430835, "clip_ratio/low_min": 0.00020395792792745246, "clip_ratio/region_mean": 0.0016489459851679082, "completions/clipped_ratio": 0.0, "completions/max_length": 517.9, "completions/max_terminated_length": 517.9, "completions/mean_length": 204.2182357788086, "completions/mean_terminated_length": 204.2182357788086, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.02565524500484268, "epoch": 19.718702173247628, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.006256103515625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3297403680.0, "reward": 7.922868013381958, "reward_std": 0.29611380994319914, "rewards/combined_reward/mean": 7.9228675842285154, "rewards/combined_reward/std": 2.392642021179199, "step": 32220 }, { "clip_ratio/high_max": 0.0021328337437201603, "clip_ratio/high_mean": 0.001127438634406038, "clip_ratio/low_mean": 0.0006918817009136547, "clip_ratio/low_min": 0.00019526628821040504, "clip_ratio/region_mean": 0.0018193203330156393, "completions/clipped_ratio": 0.0, "completions/max_length": 585.6, "completions/max_terminated_length": 585.6, "completions/mean_length": 198.47422485351564, "completions/mean_terminated_length": 198.47422485351564, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.02789290410776933, "epoch": 19.737067646158554, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3300465981.0, "reward": 7.998132181167603, "reward_std": 0.3388523399829865, "rewards/combined_reward/mean": 7.998131895065308, "rewards/combined_reward/std": 2.43163423538208, "step": 32250 }, { "clip_ratio/high_max": 0.0021427093481179325, "clip_ratio/high_mean": 0.001315791438779949, "clip_ratio/low_mean": 0.0007332150812241404, "clip_ratio/low_min": 0.00022321572275056192, "clip_ratio/region_mean": 0.0020490065280076427, "completions/clipped_ratio": 0.0, "completions/max_length": 571.6, "completions/max_terminated_length": 571.6, "completions/mean_length": 204.1916732788086, "completions/mean_terminated_length": 204.1916732788086, "completions/min_length": 51.9, "completions/min_terminated_length": 51.9, "entropy": 0.030148717357466617, "epoch": 19.755433119069483, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0264892578125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3303550237.0, "reward": 8.055132627487183, "reward_std": 0.353509946167469, "rewards/combined_reward/mean": 8.055132293701172, "rewards/combined_reward/std": 2.313204836845398, "step": 32280 }, { "clip_ratio/high_max": 0.0014014880260219797, "clip_ratio/high_mean": 0.0008018260339061575, "clip_ratio/low_mean": 0.0005972578732325928, "clip_ratio/low_min": 0.00016004161176776203, "clip_ratio/region_mean": 0.001399083902166846, "completions/clipped_ratio": 0.0, "completions/max_length": 666.8, "completions/max_terminated_length": 666.8, "completions/mean_length": 195.9210998535156, "completions/mean_terminated_length": 195.9210998535156, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.024711764751312634, "epoch": 19.77379859198041, "frac_reward_zero_std": 0.2291666731238365, "grad_norm": 0.00518798828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3306602734.0, "reward": 8.389469337463378, "reward_std": 0.26358153373003007, "rewards/combined_reward/mean": 8.389468908309937, "rewards/combined_reward/std": 1.9598036885261536, "step": 32310 }, { "clip_ratio/high_max": 0.001762258503489041, "clip_ratio/high_mean": 0.0009959628191912391, "clip_ratio/low_mean": 0.0007487069829949178, "clip_ratio/low_min": 0.00020907684205061136, "clip_ratio/region_mean": 0.001744669796122859, "completions/clipped_ratio": 0.0, "completions/max_length": 551.8, "completions/max_terminated_length": 551.8, "completions/mean_length": 207.30989990234374, "completions/mean_terminated_length": 207.30989990234374, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.028090007998980582, "epoch": 19.792164064891338, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3309698964.0, "reward": 8.202644634246827, "reward_std": 0.31000567227602005, "rewards/combined_reward/mean": 8.20264458656311, "rewards/combined_reward/std": 2.174365985393524, "step": 32340 }, { "clip_ratio/high_max": 0.002134632991510443, "clip_ratio/high_mean": 0.0012969667577029517, "clip_ratio/low_mean": 0.0007377803503307707, "clip_ratio/low_min": 0.00018610793049447238, "clip_ratio/region_mean": 0.0020347471009396637, "completions/clipped_ratio": 0.0, "completions/max_length": 673.9, "completions/max_terminated_length": 673.9, "completions/mean_length": 201.4987045288086, "completions/mean_terminated_length": 201.4987045288086, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.030155469244346023, "epoch": 19.810529537802264, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.0279541015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3312772879.0, "reward": 7.82008376121521, "reward_std": 0.3284981235861778, "rewards/combined_reward/mean": 7.8200836181640625, "rewards/combined_reward/std": 2.4721391797065735, "step": 32370 }, { "clip_ratio/high_max": 0.0022878990714768105, "clip_ratio/high_mean": 0.0013666846874305822, "clip_ratio/low_mean": 0.0007300959526522396, "clip_ratio/low_min": 0.0001806089802509329, "clip_ratio/region_mean": 0.0020967806519668858, "completions/clipped_ratio": 0.0, "completions/max_length": 594.5, "completions/max_terminated_length": 594.5, "completions/mean_length": 204.08437957763672, "completions/mean_terminated_length": 204.08437957763672, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03273746071693798, "epoch": 19.828895010713193, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.0546875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3315856723.0, "reward": 7.962983226776123, "reward_std": 0.40533367097377776, "rewards/combined_reward/mean": 7.962982559204102, "rewards/combined_reward/std": 2.2569504499435427, "step": 32400 }, { "clip_ratio/high_max": 0.0017530035448241202, "clip_ratio/high_mean": 0.0009920203569587708, "clip_ratio/low_mean": 0.000645574098113381, "clip_ratio/low_min": 0.0001538021567588051, "clip_ratio/region_mean": 0.0016375944416116303, "completions/clipped_ratio": 0.0, "completions/max_length": 490.7, "completions/max_terminated_length": 490.7, "completions/mean_length": 185.91041870117186, "completions/mean_terminated_length": 185.91041870117186, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.024782265598575274, "epoch": 19.84726048362412, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3318870779.0, "reward": 7.9797751903533936, "reward_std": 0.2750780276954174, "rewards/combined_reward/mean": 7.97977499961853, "rewards/combined_reward/std": 2.3411678314208983, "step": 32430 }, { "clip_ratio/high_max": 0.0018257647403515874, "clip_ratio/high_mean": 0.0010608761961824105, "clip_ratio/low_mean": 0.0006144695946204593, "clip_ratio/low_min": 0.00016045073765174795, "clip_ratio/region_mean": 0.0016753457806771622, "completions/clipped_ratio": 0.0, "completions/max_length": 579.3, "completions/max_terminated_length": 579.3, "completions/mean_length": 202.8315170288086, "completions/mean_terminated_length": 202.8315170288086, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.025606486185764273, "epoch": 19.865625956535048, "frac_reward_zero_std": 0.17083333991467953, "grad_norm": 0.0250244140625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3321949812.0, "reward": 8.084955739974976, "reward_std": 0.3082599304616451, "rewards/combined_reward/mean": 8.084955263137818, "rewards/combined_reward/std": 2.406928849220276, "step": 32460 }, { "clip_ratio/high_max": 0.0022010883689896825, "clip_ratio/high_mean": 0.0012882153396882737, "clip_ratio/low_mean": 0.0008437844841561551, "clip_ratio/low_min": 0.00017209034582871635, "clip_ratio/region_mean": 0.0021319998384569773, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 207.60052642822265, "completions/mean_terminated_length": 207.60052642822265, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03328916885269185, "epoch": 19.883991429445974, "frac_reward_zero_std": 0.13750000149011612, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 3325047158.0, "reward": 8.104700613021851, "reward_std": 0.42325415313243864, "rewards/combined_reward/mean": 8.104700374603272, "rewards/combined_reward/std": 2.234854817390442, "step": 32490 }, { "clip_ratio/high_max": 0.0019207540521165357, "clip_ratio/high_mean": 0.0011017411819921109, "clip_ratio/low_mean": 0.0009148193297733087, "clip_ratio/low_min": 0.00024882757061277515, "clip_ratio/region_mean": 0.0020165604966071744, "completions/clipped_ratio": 0.0, "completions/max_length": 600.8, "completions/max_terminated_length": 600.8, "completions/mean_length": 205.33333892822264, "completions/mean_terminated_length": 205.33333892822264, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.03101524943485856, "epoch": 19.902356902356903, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.006439208984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3328135798.0, "reward": 8.115864801406861, "reward_std": 0.28042294532060624, "rewards/combined_reward/mean": 8.115864324569703, "rewards/combined_reward/std": 2.1356717705726624, "step": 32520 }, { "clip_ratio/high_max": 0.0019672019804905477, "clip_ratio/high_mean": 0.0010606817435473205, "clip_ratio/low_mean": 0.0008317187028296757, "clip_ratio/low_min": 0.000278566586348461, "clip_ratio/region_mean": 0.0018924004572909326, "completions/clipped_ratio": 0.0, "completions/max_length": 617.6, "completions/max_terminated_length": 617.6, "completions/mean_length": 211.2335983276367, "completions/mean_terminated_length": 211.2335983276367, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.0311668173254778, "epoch": 19.92072237526783, "frac_reward_zero_std": 0.12083333507180213, "grad_norm": 0.005950927734375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3331247095.0, "reward": 7.748036003112793, "reward_std": 0.3424113497138023, "rewards/combined_reward/mean": 7.748035907745361, "rewards/combined_reward/std": 2.3995853662490845, "step": 32550 }, { "clip_ratio/high_max": 0.0019097958798132216, "clip_ratio/high_mean": 0.0010501538588869152, "clip_ratio/low_mean": 0.0006781471514841541, "clip_ratio/low_min": 0.00015304244467794585, "clip_ratio/region_mean": 0.001728301008794612, "completions/clipped_ratio": 0.0, "completions/max_length": 520.6, "completions/max_terminated_length": 520.6, "completions/mean_length": 191.54792022705078, "completions/mean_terminated_length": 191.54792022705078, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027372012473642827, "epoch": 19.939087848178758, "frac_reward_zero_std": 0.20000000707805157, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3334282799.0, "reward": 8.295022010803223, "reward_std": 0.25613425970077514, "rewards/combined_reward/mean": 8.295021629333496, "rewards/combined_reward/std": 1.8260868549346925, "step": 32580 }, { "clip_ratio/high_max": 0.0016701590837328694, "clip_ratio/high_mean": 0.0009927695035003125, "clip_ratio/low_mean": 0.0007775828416924924, "clip_ratio/low_min": 0.0003114169453814005, "clip_ratio/region_mean": 0.0017703523413122942, "completions/clipped_ratio": 0.0, "completions/max_length": 583.9, "completions/max_terminated_length": 583.9, "completions/mean_length": 218.4976608276367, "completions/mean_terminated_length": 218.4976608276367, "completions/min_length": 51.2, "completions/min_terminated_length": 51.2, "entropy": 0.026982279618581136, "epoch": 19.957453321089684, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.0308837890625, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 3337421990.0, "reward": 7.871974134445191, "reward_std": 0.40074070990085603, "rewards/combined_reward/mean": 7.871974039077759, "rewards/combined_reward/std": 2.470160496234894, "step": 32610 }, { "clip_ratio/high_max": 0.0018912574494606815, "clip_ratio/high_mean": 0.001091735403315397, "clip_ratio/low_mean": 0.0007540190569973978, "clip_ratio/low_min": 0.000184656928225498, "clip_ratio/region_mean": 0.0018457544545526616, "completions/clipped_ratio": 0.0, "completions/max_length": 571.8, "completions/max_terminated_length": 571.8, "completions/mean_length": 195.55182800292968, "completions/mean_terminated_length": 195.55182800292968, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.0295046908625712, "epoch": 19.975818794000613, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3340473069.0, "reward": 8.256044483184814, "reward_std": 0.2877568960189819, "rewards/combined_reward/mean": 8.25604419708252, "rewards/combined_reward/std": 1.9704054117202758, "step": 32640 }, { "clip_ratio/high_max": 0.0016854659106077937, "clip_ratio/high_mean": 0.0009252396250910048, "clip_ratio/low_mean": 0.0007346104774721122, "clip_ratio/low_min": 0.00012279600099039575, "clip_ratio/region_mean": 0.0016598501038970425, "completions/clipped_ratio": 0.0, "completions/max_length": 634.2, "completions/max_terminated_length": 634.2, "completions/mean_length": 205.1578170776367, "completions/mean_terminated_length": 205.1578170776367, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.028068154685509702, "epoch": 19.99418426691154, "frac_reward_zero_std": 0.14583333693444728, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3343561035.0, "reward": 8.098686170578002, "reward_std": 0.3148917257785797, "rewards/combined_reward/mean": 8.098686075210571, "rewards/combined_reward/std": 2.208976697921753, "step": 32670 }, { "clip_ratio/high_max": 0.001852982255275224, "clip_ratio/high_mean": 0.001052327255134425, "clip_ratio/low_mean": 0.000734843885453063, "clip_ratio/low_min": 0.00014866057608742267, "clip_ratio/region_mean": 0.0017871711418399067, "completions/clipped_ratio": 0.0, "completions/max_length": 629.1818181818181, "completions/max_terminated_length": 629.1818181818181, "completions/mean_length": 202.33286354758522, "completions/mean_terminated_length": 202.33286354758522, "completions/min_length": 39.81818181818182, "completions/min_terminated_length": 39.81818181818182, "entropy": 0.02537849450819805, "epoch": 20.012855831037648, "frac_reward_zero_std": 0.1628787907009775, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3346945865.0, "reward": 8.07837217504328, "reward_std": 0.31856546618721704, "rewards/combined_reward/mean": 8.078371741554953, "rewards/combined_reward/std": 2.201308488845825, "step": 32700 }, { "clip_ratio/high_max": 0.0023787925902676458, "clip_ratio/high_mean": 0.0014355550001103742, "clip_ratio/low_mean": 0.0010765631628601112, "clip_ratio/low_min": 0.00036266904935473577, "clip_ratio/region_mean": 0.002512118161151496, "completions/clipped_ratio": 0.0, "completions/max_length": 602.1, "completions/max_terminated_length": 602.1, "completions/mean_length": 205.40625610351563, "completions/mean_terminated_length": 205.40625610351563, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030025134158010284, "epoch": 20.031221303948577, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3350034785.0, "reward": 8.218461084365845, "reward_std": 0.36503741294145586, "rewards/combined_reward/mean": 8.218460893630981, "rewards/combined_reward/std": 2.146009373664856, "step": 32730 }, { "clip_ratio/high_max": 0.00220803771323214, "clip_ratio/high_mean": 0.0012979047569388058, "clip_ratio/low_mean": 0.000776406059837124, "clip_ratio/low_min": 0.00013242586380025992, "clip_ratio/region_mean": 0.0020743108175035256, "completions/clipped_ratio": 0.0, "completions/max_length": 540.4, "completions/max_terminated_length": 540.4, "completions/mean_length": 198.53203735351562, "completions/mean_terminated_length": 198.53203735351562, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.02616320732049644, "epoch": 20.049586776859503, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.01190185546875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3353097308.0, "reward": 7.9750974655151365, "reward_std": 0.39409870803356173, "rewards/combined_reward/mean": 7.975097274780273, "rewards/combined_reward/std": 2.289778447151184, "step": 32760 }, { "clip_ratio/high_max": 0.0023305963589033735, "clip_ratio/high_mean": 0.0013252369627783386, "clip_ratio/low_mean": 0.0009346989177477856, "clip_ratio/low_min": 0.000277045126373802, "clip_ratio/region_mean": 0.0022599358863468902, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 197.59115142822264, "completions/mean_terminated_length": 197.59115142822264, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.0278956629956762, "epoch": 20.067952249770432, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3356156218.0, "reward": 8.1525315284729, "reward_std": 0.26849527955055236, "rewards/combined_reward/mean": 8.152531290054322, "rewards/combined_reward/std": 2.215410220623016, "step": 32790 }, { "clip_ratio/high_max": 0.001902477825206006, "clip_ratio/high_mean": 0.0010758526498345115, "clip_ratio/low_mean": 0.0008604374788167964, "clip_ratio/low_min": 0.00020988863689126448, "clip_ratio/region_mean": 0.0019362901199201587, "completions/clipped_ratio": 0.0, "completions/max_length": 576.4, "completions/max_terminated_length": 576.4, "completions/mean_length": 208.8843795776367, "completions/mean_terminated_length": 208.8843795776367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.025324976195891698, "epoch": 20.086317722681358, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.024658203125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3359258494.0, "reward": 8.070201587677001, "reward_std": 0.3801767960190773, "rewards/combined_reward/mean": 8.070201396942139, "rewards/combined_reward/std": 2.255971610546112, "step": 32820 }, { "clip_ratio/high_max": 0.002131556956252704, "clip_ratio/high_mean": 0.0013069886803956858, "clip_ratio/low_mean": 0.0009196244111687216, "clip_ratio/low_min": 0.0002804699834086932, "clip_ratio/region_mean": 0.002226613084106551, "completions/clipped_ratio": 0.0, "completions/max_length": 697.9, "completions/max_terminated_length": 697.9, "completions/mean_length": 208.8502670288086, "completions/mean_terminated_length": 208.8502670288086, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.030283528457706172, "epoch": 20.104683195592287, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.00701904296875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3362360639.0, "reward": 8.170887851715088, "reward_std": 0.29715517684817316, "rewards/combined_reward/mean": 8.170887660980224, "rewards/combined_reward/std": 2.2525489926338196, "step": 32850 }, { "clip_ratio/high_max": 0.0024700070435452895, "clip_ratio/high_mean": 0.0014173478237353265, "clip_ratio/low_mean": 0.0008829671038256493, "clip_ratio/low_min": 0.0002567261855195587, "clip_ratio/region_mean": 0.0023003149589688595, "completions/clipped_ratio": 0.0, "completions/max_length": 579.7, "completions/max_terminated_length": 579.7, "completions/mean_length": 205.27578582763672, "completions/mean_terminated_length": 205.27578582763672, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.031003212882205845, "epoch": 20.123048668503213, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.0045166015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3365449058.0, "reward": 8.170474863052368, "reward_std": 0.30976987779140475, "rewards/combined_reward/mean": 8.170474672317505, "rewards/combined_reward/std": 2.087770903110504, "step": 32880 }, { "clip_ratio/high_max": 0.0020951817171104875, "clip_ratio/high_mean": 0.0012384251379747486, "clip_ratio/low_mean": 0.0009464449456572765, "clip_ratio/low_min": 0.0002670519214007072, "clip_ratio/region_mean": 0.0021848700862998763, "completions/clipped_ratio": 0.0, "completions/max_length": 556.6, "completions/max_terminated_length": 556.6, "completions/mean_length": 205.1554748535156, "completions/mean_terminated_length": 205.1554748535156, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02748293735397359, "epoch": 20.141414141414142, "frac_reward_zero_std": 0.16666666902601718, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3368537015.0, "reward": 8.017110061645507, "reward_std": 0.3464834153652191, "rewards/combined_reward/mean": 8.017110061645507, "rewards/combined_reward/std": 2.337067699432373, "step": 32910 }, { "clip_ratio/high_max": 0.0027528306736106364, "clip_ratio/high_mean": 0.0015631984996919831, "clip_ratio/low_mean": 0.0010158017413535467, "clip_ratio/low_min": 0.0003904686503422757, "clip_ratio/region_mean": 0.0025790002352247634, "completions/clipped_ratio": 0.0, "completions/max_length": 536.3, "completions/max_terminated_length": 536.3, "completions/mean_length": 206.01537170410157, "completions/mean_terminated_length": 206.01537170410157, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.03171523914982875, "epoch": 20.159779614325068, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3371628274.0, "reward": 8.313554906845093, "reward_std": 0.32344546020030973, "rewards/combined_reward/mean": 8.313554430007935, "rewards/combined_reward/std": 1.9568333864212035, "step": 32940 }, { "clip_ratio/high_max": 0.0022898532860078074, "clip_ratio/high_mean": 0.0013413041825212227, "clip_ratio/low_mean": 0.0008228941276077725, "clip_ratio/low_min": 0.00020981440417623768, "clip_ratio/region_mean": 0.002164198298608729, "completions/clipped_ratio": 0.0, "completions/max_length": 581.2, "completions/max_terminated_length": 581.2, "completions/mean_length": 205.60208740234376, "completions/mean_terminated_length": 205.60208740234376, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.027059814038996894, "epoch": 20.178145087235997, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3374717946.0, "reward": 8.168800258636475, "reward_std": 0.3013667553663254, "rewards/combined_reward/mean": 8.168799877166748, "rewards/combined_reward/std": 2.125939893722534, "step": 32970 }, { "clip_ratio/high_max": 0.0025225602328040014, "clip_ratio/high_mean": 0.0014299821654276458, "clip_ratio/low_mean": 0.0009662822586202917, "clip_ratio/low_min": 0.00023444665542532068, "clip_ratio/region_mean": 0.0023962644195610967, "completions/clipped_ratio": 0.0, "completions/max_length": 600.9, "completions/max_terminated_length": 600.9, "completions/mean_length": 187.65260620117186, "completions/mean_terminated_length": 187.65260620117186, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.0307171485076348, "epoch": 20.196510560146923, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.006622314453125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3377738692.0, "reward": 8.189899158477782, "reward_std": 0.269562129676342, "rewards/combined_reward/mean": 8.189898777008057, "rewards/combined_reward/std": 2.2207222104072573, "step": 33000 }, { "clip_ratio/high_max": 0.00188882362684429, "clip_ratio/high_mean": 0.0011204614414358123, "clip_ratio/low_mean": 0.0007961043892161494, "clip_ratio/low_min": 0.0001937086043957, "clip_ratio/region_mean": 0.0019165658256194245, "completions/clipped_ratio": 0.0, "completions/max_length": 670.5, "completions/max_terminated_length": 670.5, "completions/mean_length": 204.9171920776367, "completions/mean_terminated_length": 204.9171920776367, "completions/min_length": 47.4, "completions/min_terminated_length": 47.4, "entropy": 0.02569785077745716, "epoch": 20.214876033057852, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3380825734.0, "reward": 8.112857437133789, "reward_std": 0.3308491513133049, "rewards/combined_reward/mean": 8.112857341766357, "rewards/combined_reward/std": 2.2468063473701476, "step": 33030 }, { "clip_ratio/high_max": 0.0020996575466900444, "clip_ratio/high_mean": 0.0011426563066682623, "clip_ratio/low_mean": 0.0008319201851312148, "clip_ratio/low_min": 0.00024463706310295187, "clip_ratio/region_mean": 0.001974576489495424, "completions/clipped_ratio": 0.0, "completions/max_length": 600.3, "completions/max_terminated_length": 600.3, "completions/mean_length": 194.45833740234374, "completions/mean_terminated_length": 194.45833740234374, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.026141522654021778, "epoch": 20.233241505968778, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3383872614.0, "reward": 7.901655769348144, "reward_std": 0.3609850510954857, "rewards/combined_reward/mean": 7.901655578613282, "rewards/combined_reward/std": 2.378695583343506, "step": 33060 }, { "clip_ratio/high_max": 0.002006689063758434, "clip_ratio/high_mean": 0.0011527189182136985, "clip_ratio/low_mean": 0.0009051319331774721, "clip_ratio/low_min": 0.00022168165354135755, "clip_ratio/region_mean": 0.0020578508614562453, "completions/clipped_ratio": 0.0, "completions/max_length": 585.2, "completions/max_terminated_length": 585.2, "completions/mean_length": 197.52317962646484, "completions/mean_terminated_length": 197.52317962646484, "completions/min_length": 46.9, "completions/min_terminated_length": 46.9, "entropy": 0.026638616411946713, "epoch": 20.251606978879707, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3386931263.0, "reward": 8.080444765090942, "reward_std": 0.3084380075335503, "rewards/combined_reward/mean": 8.080444622039796, "rewards/combined_reward/std": 2.2385523915290833, "step": 33090 }, { "clip_ratio/high_max": 0.002694760409455436, "clip_ratio/high_mean": 0.0014685242807900068, "clip_ratio/low_mean": 0.0010678610124159604, "clip_ratio/low_min": 0.0003020348851957048, "clip_ratio/region_mean": 0.0025363853104257337, "completions/clipped_ratio": 0.0, "completions/max_length": 640.4, "completions/max_terminated_length": 640.4, "completions/mean_length": 207.7971420288086, "completions/mean_terminated_length": 207.7971420288086, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.032030554752176005, "epoch": 20.269972451790633, "frac_reward_zero_std": 0.15000000298023225, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3390029364.0, "reward": 8.353310775756835, "reward_std": 0.34732191562652587, "rewards/combined_reward/mean": 8.353310537338256, "rewards/combined_reward/std": 1.8524880409240723, "step": 33120 }, { "clip_ratio/high_max": 0.002302641180479744, "clip_ratio/high_mean": 0.0012566663911760163, "clip_ratio/low_mean": 0.0007684880726931927, "clip_ratio/low_min": 0.00012054572216584347, "clip_ratio/region_mean": 0.00202515446144389, "completions/clipped_ratio": 0.0, "completions/max_length": 571.5, "completions/max_terminated_length": 571.5, "completions/mean_length": 191.1671936035156, "completions/mean_terminated_length": 191.1671936035156, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.026808736504366, "epoch": 20.288337924701562, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3393063606.0, "reward": 7.873023939132691, "reward_std": 0.35137041807174685, "rewards/combined_reward/mean": 7.873023748397827, "rewards/combined_reward/std": 2.451361322402954, "step": 33150 }, { "clip_ratio/high_max": 0.0024513742556640256, "clip_ratio/high_mean": 0.001355385199591789, "clip_ratio/low_mean": 0.0009556719346922667, "clip_ratio/low_min": 0.0002568451001328261, "clip_ratio/region_mean": 0.002311057118155683, "completions/clipped_ratio": 0.0, "completions/max_length": 594.3, "completions/max_terminated_length": 594.3, "completions/mean_length": 193.89818115234374, "completions/mean_terminated_length": 193.89818115234374, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.03323134693006675, "epoch": 20.306703397612488, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3396108335.0, "reward": 8.135394859313966, "reward_std": 0.3224275231361389, "rewards/combined_reward/mean": 8.135394763946532, "rewards/combined_reward/std": 2.2907402992248533, "step": 33180 }, { "clip_ratio/high_max": 0.0028052085525511455, "clip_ratio/high_mean": 0.001554954947520552, "clip_ratio/low_mean": 0.0009260642424730273, "clip_ratio/low_min": 0.00023844559788509893, "clip_ratio/region_mean": 0.0024810191762905257, "completions/clipped_ratio": 0.0, "completions/max_length": 654.8, "completions/max_terminated_length": 654.8, "completions/mean_length": 199.83802795410156, "completions/mean_terminated_length": 199.83802795410156, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.033194410257662336, "epoch": 20.325068870523417, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.02587890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3399175873.0, "reward": 7.959139585494995, "reward_std": 0.28822345584630965, "rewards/combined_reward/mean": 7.959139347076416, "rewards/combined_reward/std": 2.3833719849586488, "step": 33210 }, { "clip_ratio/high_max": 0.0022797698193850616, "clip_ratio/high_mean": 0.0013266134226190237, "clip_ratio/low_mean": 0.0007656883287078623, "clip_ratio/low_min": 0.00012574440819056084, "clip_ratio/region_mean": 0.0020923017432020667, "completions/clipped_ratio": 0.0, "completions/max_length": 527.8, "completions/max_terminated_length": 527.8, "completions/mean_length": 192.0541732788086, "completions/mean_terminated_length": 192.0541732788086, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.026926339759180944, "epoch": 20.343434343434343, "frac_reward_zero_std": 0.2083333373069763, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3402213521.0, "reward": 8.241342067718506, "reward_std": 0.3090029016137123, "rewards/combined_reward/mean": 8.241341686248779, "rewards/combined_reward/std": 2.038411557674408, "step": 33240 }, { "clip_ratio/high_max": 0.002721478258899879, "clip_ratio/high_mean": 0.0015164873055861486, "clip_ratio/low_mean": 0.0009361600834381534, "clip_ratio/low_min": 0.00027967624919256195, "clip_ratio/region_mean": 0.002452647370713142, "completions/clipped_ratio": 0.0, "completions/max_length": 633.4, "completions/max_terminated_length": 633.4, "completions/mean_length": 199.21771392822265, "completions/mean_terminated_length": 199.21771392822265, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029453575727529824, "epoch": 20.361799816345272, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.004791259765625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3405278677.0, "reward": 8.276958227157593, "reward_std": 0.27630086839199064, "rewards/combined_reward/mean": 8.276957988739014, "rewards/combined_reward/std": 2.023733365535736, "step": 33270 }, { "clip_ratio/high_max": 0.0025134881027042864, "clip_ratio/high_mean": 0.0014321050388389268, "clip_ratio/low_mean": 0.0010386791323981014, "clip_ratio/low_min": 0.0003499734397337306, "clip_ratio/region_mean": 0.002470784181302103, "completions/clipped_ratio": 0.0, "completions/max_length": 530.1, "completions/max_terminated_length": 530.1, "completions/mean_length": 196.4237060546875, "completions/mean_terminated_length": 196.4237060546875, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02993291602469981, "epoch": 20.380165289256198, "frac_reward_zero_std": 0.20000000707805157, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3408333104.0, "reward": 8.185216188430786, "reward_std": 0.3406363636255264, "rewards/combined_reward/mean": 8.185215997695924, "rewards/combined_reward/std": 2.13793283700943, "step": 33300 }, { "clip_ratio/high_max": 0.0021973800350679083, "clip_ratio/high_mean": 0.0012762428951342978, "clip_ratio/low_mean": 0.0007576472889923025, "clip_ratio/low_min": 0.0001773054854614505, "clip_ratio/region_mean": 0.0020338901803673556, "completions/clipped_ratio": 0.0, "completions/max_length": 528.2, "completions/max_terminated_length": 528.2, "completions/mean_length": 191.16172485351564, "completions/mean_terminated_length": 191.16172485351564, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.025631716552500922, "epoch": 20.398530762167127, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3411367325.0, "reward": 8.442902946472168, "reward_std": 0.2819958031177521, "rewards/combined_reward/mean": 8.442902612686158, "rewards/combined_reward/std": 1.824765384197235, "step": 33330 }, { "clip_ratio/high_max": 0.0024424345479928888, "clip_ratio/high_mean": 0.0014103732391959056, "clip_ratio/low_mean": 0.000992742640422269, "clip_ratio/low_min": 0.00028623306352528746, "clip_ratio/region_mean": 0.0024031159041138987, "completions/clipped_ratio": 0.0, "completions/max_length": 595.4, "completions/max_terminated_length": 595.4, "completions/mean_length": 194.5104202270508, "completions/mean_terminated_length": 194.5104202270508, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02653946642143031, "epoch": 20.416896235078053, "frac_reward_zero_std": 0.18750000782310963, "grad_norm": 0.00439453125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3414414405.0, "reward": 7.914232873916626, "reward_std": 0.2921629935503006, "rewards/combined_reward/mean": 7.914232921600342, "rewards/combined_reward/std": 2.3917328000068663, "step": 33360 }, { "clip_ratio/high_max": 0.0021423461611751312, "clip_ratio/high_mean": 0.0011816627042207984, "clip_ratio/low_mean": 0.0009167591694373792, "clip_ratio/low_min": 0.00027610045168936873, "clip_ratio/region_mean": 0.0020984218712934914, "completions/clipped_ratio": 0.0, "completions/max_length": 577.6, "completions/max_terminated_length": 577.6, "completions/mean_length": 194.81458740234376, "completions/mean_terminated_length": 194.81458740234376, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029085091035813092, "epoch": 20.435261707988982, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3417462653.0, "reward": 8.027475547790527, "reward_std": 0.34967841804027555, "rewards/combined_reward/mean": 8.027475357055664, "rewards/combined_reward/std": 2.4236528873443604, "step": 33390 }, { "clip_ratio/high_max": 0.0022825633830507288, "clip_ratio/high_mean": 0.0013023138840556688, "clip_ratio/low_mean": 0.0009447689021423382, "clip_ratio/low_min": 0.0002765048043026278, "clip_ratio/region_mean": 0.002247082791776241, "completions/clipped_ratio": 0.0, "completions/max_length": 518.2, "completions/max_terminated_length": 518.2, "completions/mean_length": 194.10442962646485, "completions/mean_terminated_length": 194.10442962646485, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.02781227285352846, "epoch": 20.453627180899908, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3420508174.0, "reward": 7.881655979156494, "reward_std": 0.34899852573871615, "rewards/combined_reward/mean": 7.881655550003051, "rewards/combined_reward/std": 2.4803265929222107, "step": 33420 }, { "clip_ratio/high_max": 0.002500199504235449, "clip_ratio/high_mean": 0.001358687015454052, "clip_ratio/low_mean": 0.0008656039800068054, "clip_ratio/low_min": 0.0002401234603894409, "clip_ratio/region_mean": 0.002224290989882623, "completions/clipped_ratio": 0.0, "completions/max_length": 591.9, "completions/max_terminated_length": 591.9, "completions/mean_length": 205.83698577880858, "completions/mean_terminated_length": 205.83698577880858, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.027179854076045253, "epoch": 20.471992653810837, "frac_reward_zero_std": 0.1291666693985462, "grad_norm": 0.008056640625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3423598748.0, "reward": 8.118167114257812, "reward_std": 0.3400725156068802, "rewards/combined_reward/mean": 8.11816701889038, "rewards/combined_reward/std": 2.3761266589164736, "step": 33450 }, { "clip_ratio/high_max": 0.0025444538706021072, "clip_ratio/high_mean": 0.0013320095549715916, "clip_ratio/low_mean": 0.0009116920655894016, "clip_ratio/low_min": 0.000241291237277134, "clip_ratio/region_mean": 0.0022437016220161845, "completions/clipped_ratio": 0.0, "completions/max_length": 527.1, "completions/max_terminated_length": 527.1, "completions/mean_length": 184.30234985351564, "completions/mean_terminated_length": 184.30234985351564, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.027881667669862508, "epoch": 20.490358126721762, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3426606629.0, "reward": 8.155625534057616, "reward_std": 0.3241216570138931, "rewards/combined_reward/mean": 8.155625009536744, "rewards/combined_reward/std": 2.2586544036865233, "step": 33480 }, { "clip_ratio/high_max": 0.0022420475230319425, "clip_ratio/high_mean": 0.001300808415665718, "clip_ratio/low_mean": 0.0010300527001769902, "clip_ratio/low_min": 0.0003266678688911876, "clip_ratio/region_mean": 0.002330861114023719, "completions/clipped_ratio": 0.0, "completions/max_length": 633.5, "completions/max_terminated_length": 633.5, "completions/mean_length": 193.4781295776367, "completions/mean_terminated_length": 193.4781295776367, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029418817441910505, "epoch": 20.50872359963269, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3429649745.0, "reward": 7.9963549137115475, "reward_std": 0.3251468658447266, "rewards/combined_reward/mean": 7.996354722976685, "rewards/combined_reward/std": 2.291062080860138, "step": 33510 }, { "clip_ratio/high_max": 0.0023610477175679987, "clip_ratio/high_mean": 0.0013275557133965777, "clip_ratio/low_mean": 0.0011094242547793935, "clip_ratio/low_min": 0.0002759974777291063, "clip_ratio/region_mean": 0.0024369799764826896, "completions/clipped_ratio": 0.0, "completions/max_length": 580.7, "completions/max_terminated_length": 580.7, "completions/mean_length": 198.59818115234376, "completions/mean_terminated_length": 198.59818115234376, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.034265308151952924, "epoch": 20.527089072543617, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.006988525390625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3432712522.0, "reward": 7.960395336151123, "reward_std": 0.29146433621644974, "rewards/combined_reward/mean": 7.960395097732544, "rewards/combined_reward/std": 2.4722677946090696, "step": 33540 }, { "clip_ratio/high_max": 0.00218031074279376, "clip_ratio/high_mean": 0.0011955571757425788, "clip_ratio/low_mean": 0.0009759388046707803, "clip_ratio/low_min": 0.00034878551038370156, "clip_ratio/region_mean": 0.0021714959701057525, "completions/clipped_ratio": 0.0, "completions/max_length": 605.6, "completions/max_terminated_length": 605.6, "completions/mean_length": 205.02422637939452, "completions/mean_terminated_length": 205.02422637939452, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.025935579867412645, "epoch": 20.545454545454547, "frac_reward_zero_std": 0.19583334103226663, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3435799975.0, "reward": 8.23196063041687, "reward_std": 0.33376126289367675, "rewards/combined_reward/mean": 8.231960248947143, "rewards/combined_reward/std": 2.073754584789276, "step": 33570 }, { "clip_ratio/high_max": 0.0029923128546215593, "clip_ratio/high_mean": 0.0015953370771361127, "clip_ratio/low_mean": 0.0010596436923757816, "clip_ratio/low_min": 0.00032921338788582943, "clip_ratio/region_mean": 0.002654980779819501, "completions/clipped_ratio": 0.0, "completions/max_length": 600.2, "completions/max_terminated_length": 600.2, "completions/mean_length": 194.93984985351562, "completions/mean_terminated_length": 194.93984985351562, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.02936845370568335, "epoch": 20.563820018365472, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01025390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3438848704.0, "reward": 8.133590173721313, "reward_std": 0.28114933520555496, "rewards/combined_reward/mean": 8.13359031677246, "rewards/combined_reward/std": 2.034004235267639, "step": 33600 }, { "clip_ratio/high_max": 0.002686879365258695, "clip_ratio/high_mean": 0.0015803903702665896, "clip_ratio/low_mean": 0.0008258109785553339, "clip_ratio/low_min": 0.00015752294154178041, "clip_ratio/region_mean": 0.002406201328752407, "completions/clipped_ratio": 0.0, "completions/max_length": 572.4, "completions/max_terminated_length": 572.4, "completions/mean_length": 206.9869842529297, "completions/mean_terminated_length": 206.9869842529297, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.036110443016514185, "epoch": 20.581573308846036, "frac_reward_zero_std": 0.14583333507180213, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3441943694.0, "reward": 8.253213167190552, "reward_std": 0.3690024107694626, "rewards/combined_reward/mean": 8.25321307182312, "rewards/combined_reward/std": 2.150123381614685, "step": 33630 }, { "clip_ratio/high_max": 0.0019987452062196097, "clip_ratio/high_mean": 0.0010862200380264162, "clip_ratio/low_mean": 0.000769205104491751, "clip_ratio/low_min": 0.0001647441369035126, "clip_ratio/region_mean": 0.001855425142760699, "completions/clipped_ratio": 0.0, "completions/max_length": 580.7, "completions/max_terminated_length": 580.7, "completions/mean_length": 211.55573425292968, "completions/mean_terminated_length": 211.55573425292968, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.02670922178464631, "epoch": 20.599938781756965, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3445056228.0, "reward": 8.170700788497925, "reward_std": 0.3348456099629402, "rewards/combined_reward/mean": 8.17070050239563, "rewards/combined_reward/std": 2.0972132802009584, "step": 33660 }, { "clip_ratio/high_max": 0.0021244342036273642, "clip_ratio/high_mean": 0.0012547930018627085, "clip_ratio/low_mean": 0.0008539216959131106, "clip_ratio/low_min": 0.00022944470789904397, "clip_ratio/region_mean": 0.0021087147023839256, "completions/clipped_ratio": 0.0, "completions/max_length": 596.1, "completions/max_terminated_length": 596.1, "completions/mean_length": 188.4718795776367, "completions/mean_terminated_length": 188.4718795776367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.030727079596060018, "epoch": 20.61830425466789, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.01556396484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3448080120.0, "reward": 7.81871280670166, "reward_std": 0.3337065249681473, "rewards/combined_reward/mean": 7.818712472915649, "rewards/combined_reward/std": 2.455910551548004, "step": 33690 }, { "clip_ratio/high_max": 0.002105805461178534, "clip_ratio/high_mean": 0.0012224209264483459, "clip_ratio/low_mean": 0.0007082981224812102, "clip_ratio/low_min": 0.000184865928410242, "clip_ratio/region_mean": 0.0019307190532951305, "completions/clipped_ratio": 0.0, "completions/max_length": 521.6, "completions/max_terminated_length": 521.6, "completions/mean_length": 199.4315155029297, "completions/mean_terminated_length": 199.4315155029297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.028403599684437116, "epoch": 20.63666972757882, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3451146097.0, "reward": 8.023822402954101, "reward_std": 0.31862876415252683, "rewards/combined_reward/mean": 8.02382197380066, "rewards/combined_reward/std": 2.274700140953064, "step": 33720 }, { "clip_ratio/high_max": 0.001915561432542745, "clip_ratio/high_mean": 0.0011205972522778515, "clip_ratio/low_mean": 0.0007630766188716128, "clip_ratio/low_min": 0.00019540114347667745, "clip_ratio/region_mean": 0.0018836738871565708, "completions/clipped_ratio": 0.0, "completions/max_length": 581.5, "completions/max_terminated_length": 581.5, "completions/mean_length": 196.49896392822265, "completions/mean_terminated_length": 196.49896392822265, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.028744777796479562, "epoch": 20.655035200489746, "frac_reward_zero_std": 0.15416666828095912, "grad_norm": 0.005126953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3454200813.0, "reward": 8.12913236618042, "reward_std": 0.30350392013788224, "rewards/combined_reward/mean": 8.129132175445557, "rewards/combined_reward/std": 2.251833951473236, "step": 33750 }, { "clip_ratio/high_max": 0.001726694377430249, "clip_ratio/high_mean": 0.001090239386879451, "clip_ratio/low_mean": 0.0007459441069537812, "clip_ratio/low_min": 0.00025233677345871307, "clip_ratio/region_mean": 0.0018361834930450034, "completions/clipped_ratio": 0.0, "completions/max_length": 629.2, "completions/max_terminated_length": 629.2, "completions/mean_length": 188.7296920776367, "completions/mean_terminated_length": 188.7296920776367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027019665374731026, "epoch": 20.673400673400675, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3457225695.0, "reward": 7.996997833251953, "reward_std": 0.3022454336285591, "rewards/combined_reward/mean": 7.99699764251709, "rewards/combined_reward/std": 2.2971241950988768, "step": 33780 }, { "clip_ratio/high_max": 0.0019557227826832485, "clip_ratio/high_mean": 0.001010750763695493, "clip_ratio/low_mean": 0.000635845400878073, "clip_ratio/low_min": 0.0001334124387843379, "clip_ratio/region_mean": 0.0016465961659074917, "completions/clipped_ratio": 0.0, "completions/max_length": 515.5, "completions/max_terminated_length": 515.5, "completions/mean_length": 186.73281707763672, "completions/mean_terminated_length": 186.73281707763672, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02891107862815261, "epoch": 20.6917661463116, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3460242909.0, "reward": 8.048765707015992, "reward_std": 0.285368288308382, "rewards/combined_reward/mean": 8.048765563964844, "rewards/combined_reward/std": 2.2764281034469604, "step": 33810 }, { "clip_ratio/high_max": 0.0018930889292581318, "clip_ratio/high_mean": 0.001120018605312604, "clip_ratio/low_mean": 0.0007905120814636272, "clip_ratio/low_min": 0.0002223710497977057, "clip_ratio/region_mean": 0.001910530669071401, "completions/clipped_ratio": 0.0, "completions/max_length": 593.6, "completions/max_terminated_length": 593.6, "completions/mean_length": 206.16927490234374, "completions/mean_terminated_length": 206.16927490234374, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.031750053027644755, "epoch": 20.71013161922253, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.03125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3463334759.0, "reward": 7.9958759307861325, "reward_std": 0.3641438752412796, "rewards/combined_reward/mean": 7.9958757877349855, "rewards/combined_reward/std": 2.448261082172394, "step": 33840 }, { "clip_ratio/high_max": 0.0015041005324746947, "clip_ratio/high_mean": 0.0009118941054111929, "clip_ratio/low_mean": 0.0007283446519674423, "clip_ratio/low_min": 0.00021050225865716735, "clip_ratio/region_mean": 0.0016402387749015663, "completions/clipped_ratio": 0.0, "completions/max_length": 554.2, "completions/max_terminated_length": 554.2, "completions/mean_length": 194.53073272705078, "completions/mean_terminated_length": 194.53073272705078, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.0266140629692624, "epoch": 20.728497092133455, "frac_reward_zero_std": 0.13333333656191826, "grad_norm": 0.004547119140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3466381917.0, "reward": 8.093869352340699, "reward_std": 0.29405568838119506, "rewards/combined_reward/mean": 8.093869256973267, "rewards/combined_reward/std": 2.2299705862998964, "step": 33870 }, { "clip_ratio/high_max": 0.0019305577035993338, "clip_ratio/high_mean": 0.001128393782467659, "clip_ratio/low_mean": 0.0007778314289074236, "clip_ratio/low_min": 0.00020278446390875615, "clip_ratio/region_mean": 0.0019062252079796356, "completions/clipped_ratio": 0.0, "completions/max_length": 530.7, "completions/max_terminated_length": 530.7, "completions/mean_length": 198.00209045410156, "completions/mean_terminated_length": 198.00209045410156, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02699973213020712, "epoch": 20.746862565044385, "frac_reward_zero_std": 0.23750000670552254, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3469442405.0, "reward": 8.086074352264404, "reward_std": 0.2554665729403496, "rewards/combined_reward/mean": 8.086073970794677, "rewards/combined_reward/std": 2.140595316886902, "step": 33900 }, { "clip_ratio/high_max": 0.0016995411759126, "clip_ratio/high_mean": 0.0009543661535038457, "clip_ratio/low_mean": 0.0008092074713204056, "clip_ratio/low_min": 0.0002463988877328423, "clip_ratio/region_mean": 0.0017635736294323578, "completions/clipped_ratio": 0.0, "completions/max_length": 581.9, "completions/max_terminated_length": 581.9, "completions/mean_length": 198.2963607788086, "completions/mean_terminated_length": 198.2963607788086, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03025532984174788, "epoch": 20.76522803795531, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.005340576171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3472504023.0, "reward": 8.267981910705567, "reward_std": 0.24014504104852677, "rewards/combined_reward/mean": 8.267981767654419, "rewards/combined_reward/std": 2.0401092410087585, "step": 33930 }, { "clip_ratio/high_max": 0.001864344338294662, "clip_ratio/high_mean": 0.0010844309047873442, "clip_ratio/low_mean": 0.0008643066434160573, "clip_ratio/low_min": 0.0002624473168301241, "clip_ratio/region_mean": 0.0019487375238289436, "completions/clipped_ratio": 0.0, "completions/max_length": 579.5, "completions/max_terminated_length": 579.5, "completions/mean_length": 190.40443267822266, "completions/mean_terminated_length": 190.40443267822266, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.028095041975999872, "epoch": 20.78359351086624, "frac_reward_zero_std": 0.18750000298023223, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3475535336.0, "reward": 8.338893556594849, "reward_std": 0.29634298458695413, "rewards/combined_reward/mean": 8.338893365859985, "rewards/combined_reward/std": 2.049321210384369, "step": 33960 }, { "clip_ratio/high_max": 0.0018070701587324341, "clip_ratio/high_mean": 0.0009656283822550904, "clip_ratio/low_mean": 0.0006514726486784639, "clip_ratio/low_min": 0.00016165986792960514, "clip_ratio/region_mean": 0.0016171010123798625, "completions/clipped_ratio": 0.0, "completions/max_length": 548.3, "completions/max_terminated_length": 548.3, "completions/mean_length": 205.14609985351564, "completions/mean_terminated_length": 205.14609985351564, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028159775972987213, "epoch": 20.801958983777165, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3478623257.0, "reward": 8.139327573776246, "reward_std": 0.34295104146003724, "rewards/combined_reward/mean": 8.139327239990234, "rewards/combined_reward/std": 2.2248833417892455, "step": 33990 }, { "clip_ratio/high_max": 0.0016994555548687155, "clip_ratio/high_mean": 0.0010162085496025004, "clip_ratio/low_mean": 0.0008102339275258903, "clip_ratio/low_min": 0.00017895911223604345, "clip_ratio/region_mean": 0.0018264424725202844, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 200.11328887939453, "completions/mean_terminated_length": 200.11328887939453, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.030011994554661213, "epoch": 20.820324456688095, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3481691852.0, "reward": 8.17683548927307, "reward_std": 0.2741904303431511, "rewards/combined_reward/mean": 8.176835107803345, "rewards/combined_reward/std": 2.149830627441406, "step": 34020 }, { "clip_ratio/high_max": 0.00219338687529671, "clip_ratio/high_mean": 0.0013001601188686134, "clip_ratio/low_mean": 0.0008885419872967759, "clip_ratio/low_min": 0.0002716334243208015, "clip_ratio/region_mean": 0.0021887021294484535, "completions/clipped_ratio": 0.0, "completions/max_length": 576.7, "completions/max_terminated_length": 576.7, "completions/mean_length": 195.00573272705077, "completions/mean_terminated_length": 195.00573272705077, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.030563113652169703, "epoch": 20.83868992959902, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.01226806640625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3484740834.0, "reward": 7.970830249786377, "reward_std": 0.33492658734321595, "rewards/combined_reward/mean": 7.970830249786377, "rewards/combined_reward/std": 2.2412051439285277, "step": 34050 }, { "clip_ratio/high_max": 0.0017229499921086245, "clip_ratio/high_mean": 0.0010911555706115904, "clip_ratio/low_mean": 0.0007468179381248774, "clip_ratio/low_min": 0.0002049754208807523, "clip_ratio/region_mean": 0.0018379735042496274, "completions/clipped_ratio": 0.0, "completions/max_length": 589.6, "completions/max_terminated_length": 589.6, "completions/mean_length": 193.55677795410156, "completions/mean_terminated_length": 193.55677795410156, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.028191712871193885, "epoch": 20.85705540250995, "frac_reward_zero_std": 0.22083333656191825, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3487784252.0, "reward": 8.203554821014404, "reward_std": 0.30434202328324317, "rewards/combined_reward/mean": 8.203554677963258, "rewards/combined_reward/std": 2.0741513848304747, "step": 34080 }, { "clip_ratio/high_max": 0.0020378891053648355, "clip_ratio/high_mean": 0.001233848711005218, "clip_ratio/low_mean": 0.000676447536291865, "clip_ratio/low_min": 0.00015524994014413097, "clip_ratio/region_mean": 0.001910296250571264, "completions/clipped_ratio": 0.0, "completions/max_length": 585.2, "completions/max_terminated_length": 585.2, "completions/mean_length": 203.9796936035156, "completions/mean_terminated_length": 203.9796936035156, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.03215597135325273, "epoch": 20.875420875420875, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3490867694.0, "reward": 8.306687927246093, "reward_std": 0.38889536410570147, "rewards/combined_reward/mean": 8.306687498092652, "rewards/combined_reward/std": 1.9981181979179383, "step": 34110 }, { "clip_ratio/high_max": 0.0021543072197043028, "clip_ratio/high_mean": 0.0012183650163933636, "clip_ratio/low_mean": 0.0007213793869595975, "clip_ratio/low_min": 0.00019628694620526705, "clip_ratio/region_mean": 0.001939744430516536, "completions/clipped_ratio": 0.0, "completions/max_length": 574.6, "completions/max_terminated_length": 574.6, "completions/mean_length": 192.20547180175782, "completions/mean_terminated_length": 192.20547180175782, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.03051832804145912, "epoch": 20.8937863483318, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3493905923.0, "reward": 8.065681409835815, "reward_std": 0.31478718370199205, "rewards/combined_reward/mean": 8.06568112373352, "rewards/combined_reward/std": 2.2783772110939027, "step": 34140 }, { "clip_ratio/high_max": 0.002048406219788982, "clip_ratio/high_mean": 0.0012083322823552105, "clip_ratio/low_mean": 0.0007211828928120667, "clip_ratio/low_min": 0.00021927826116249587, "clip_ratio/region_mean": 0.0019295151811093091, "completions/clipped_ratio": 0.0, "completions/max_length": 599.8, "completions/max_terminated_length": 599.8, "completions/mean_length": 203.22995300292968, "completions/mean_terminated_length": 203.22995300292968, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03153946824992696, "epoch": 20.91215182124273, "frac_reward_zero_std": 0.21666666939854623, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3496986486.0, "reward": 8.21604962348938, "reward_std": 0.32361159920692445, "rewards/combined_reward/mean": 8.216049337387085, "rewards/combined_reward/std": 2.1125471591949463, "step": 34170 }, { "clip_ratio/high_max": 0.002019388422195334, "clip_ratio/high_mean": 0.001205530177927964, "clip_ratio/low_mean": 0.0007176214024487612, "clip_ratio/low_min": 0.0002465533344851186, "clip_ratio/region_mean": 0.0019231515684320281, "completions/clipped_ratio": 0.0, "completions/max_length": 545.8, "completions/max_terminated_length": 545.8, "completions/mean_length": 191.1783874511719, "completions/mean_terminated_length": 191.1783874511719, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.028883973276242612, "epoch": 20.930517294153656, "frac_reward_zero_std": 0.20000000447034835, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3500020771.0, "reward": 7.802030467987061, "reward_std": 0.38464898318052293, "rewards/combined_reward/mean": 7.802030229568482, "rewards/combined_reward/std": 2.616735565662384, "step": 34200 }, { "clip_ratio/high_max": 0.0017258572159335017, "clip_ratio/high_mean": 0.0009782564375200309, "clip_ratio/low_mean": 0.0009042422940183315, "clip_ratio/low_min": 0.0003194729521055706, "clip_ratio/region_mean": 0.0018824987239592398, "completions/clipped_ratio": 0.0, "completions/max_length": 632.2, "completions/max_terminated_length": 632.2, "completions/mean_length": 191.4596420288086, "completions/mean_terminated_length": 191.4596420288086, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027859714549655714, "epoch": 20.948882767064585, "frac_reward_zero_std": 0.14583333879709243, "grad_norm": 0.02294921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3503056136.0, "reward": 7.924511957168579, "reward_std": 0.2871484249830246, "rewards/combined_reward/mean": 7.9245116233825685, "rewards/combined_reward/std": 2.238017225265503, "step": 34230 }, { "clip_ratio/high_max": 0.0018011207449793196, "clip_ratio/high_mean": 0.0010205703918472863, "clip_ratio/low_mean": 0.0008613676139551293, "clip_ratio/low_min": 0.00028084985654762326, "clip_ratio/region_mean": 0.0018819379853084683, "completions/clipped_ratio": 0.0, "completions/max_length": 577.5, "completions/max_terminated_length": 577.5, "completions/mean_length": 205.0460998535156, "completions/mean_terminated_length": 205.0460998535156, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029864499686906736, "epoch": 20.96724823997551, "frac_reward_zero_std": 0.15000000260770321, "grad_norm": 0.00677490234375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3506143673.0, "reward": 7.959806251525879, "reward_std": 0.2925576835870743, "rewards/combined_reward/mean": 7.959805870056153, "rewards/combined_reward/std": 2.3730597257614137, "step": 34260 }, { "clip_ratio/high_max": 0.002212027183365232, "clip_ratio/high_mean": 0.0011621185719074372, "clip_ratio/low_mean": 0.0009180516653335265, "clip_ratio/low_min": 0.0003052613593657346, "clip_ratio/region_mean": 0.002080170226690825, "completions/clipped_ratio": 0.0, "completions/max_length": 625.3, "completions/max_terminated_length": 625.3, "completions/mean_length": 198.47709045410156, "completions/mean_terminated_length": 198.47709045410156, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.030357154629503686, "epoch": 20.98561371288644, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3509205985.0, "reward": 7.9746057987213135, "reward_std": 0.27948084771633147, "rewards/combined_reward/mean": 7.974605417251587, "rewards/combined_reward/std": 2.3822561860084535, "step": 34290 }, { "clip_ratio/high_max": 0.0017423892353630823, "clip_ratio/high_mean": 0.0009180765952552348, "clip_ratio/low_mean": 0.00066146732390988, "clip_ratio/low_min": 0.00017851034749955793, "clip_ratio/region_mean": 0.0015795439089072074, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 201.75852688876066, "completions/mean_terminated_length": 201.75852688876066, "completions/min_length": 39.63636363636363, "completions/min_terminated_length": 39.63636363636363, "entropy": 0.02706527138189947, "epoch": 21.00428527701255, "frac_reward_zero_std": 0.13257576051083478, "grad_norm": 0.02392578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3512588389.0, "reward": 8.132137211886318, "reward_std": 0.30667182125828485, "rewards/combined_reward/mean": 8.132136995142156, "rewards/combined_reward/std": 2.146474079652266, "step": 34320 }, { "clip_ratio/high_max": 0.0028215869606356137, "clip_ratio/high_mean": 0.0015379359421785921, "clip_ratio/low_mean": 0.0008480822992472289, "clip_ratio/low_min": 0.00021053996000167292, "clip_ratio/region_mean": 0.0023860182118369266, "completions/clipped_ratio": 0.0, "completions/max_length": 554.8, "completions/max_terminated_length": 554.8, "completions/mean_length": 198.14401550292968, "completions/mean_terminated_length": 198.14401550292968, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.029328341238821547, "epoch": 21.02265074992348, "frac_reward_zero_std": 0.22500000819563865, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3515649422.0, "reward": 7.896756219863891, "reward_std": 0.2999565929174423, "rewards/combined_reward/mean": 7.896756172180176, "rewards/combined_reward/std": 2.245482158660889, "step": 34350 }, { "clip_ratio/high_max": 0.0027546667183438937, "clip_ratio/high_mean": 0.0015986530338219988, "clip_ratio/low_mean": 0.0009069350401356739, "clip_ratio/low_min": 0.000265119283964547, "clip_ratio/region_mean": 0.002505588084265279, "completions/clipped_ratio": 0.0, "completions/max_length": 572.4, "completions/max_terminated_length": 572.4, "completions/mean_length": 208.8041732788086, "completions/mean_terminated_length": 208.8041732788086, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03306100904010236, "epoch": 21.041016222834404, "frac_reward_zero_std": 0.14166666716337203, "grad_norm": 0.006439208984375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3518751390.0, "reward": 8.112790489196778, "reward_std": 0.2980217315256596, "rewards/combined_reward/mean": 8.112790203094482, "rewards/combined_reward/std": 2.343194043636322, "step": 34380 }, { "clip_ratio/high_max": 0.002506869896024, "clip_ratio/high_mean": 0.0013785081585713972, "clip_ratio/low_mean": 0.0008971509044689204, "clip_ratio/low_min": 0.00022019358148099854, "clip_ratio/region_mean": 0.002275659058068413, "completions/clipped_ratio": 0.0, "completions/max_length": 655.8, "completions/max_terminated_length": 655.8, "completions/mean_length": 204.39557952880858, "completions/mean_terminated_length": 204.39557952880858, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.027430232955763738, "epoch": 21.059381695745333, "frac_reward_zero_std": 0.16250000409781934, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3521836429.0, "reward": 7.935137176513672, "reward_std": 0.3738603383302689, "rewards/combined_reward/mean": 7.935137033462524, "rewards/combined_reward/std": 2.4922028303146364, "step": 34410 }, { "clip_ratio/high_max": 0.0027639397803189544, "clip_ratio/high_mean": 0.0017357969897905908, "clip_ratio/low_mean": 0.0010138553108845372, "clip_ratio/low_min": 0.0002450266935435745, "clip_ratio/region_mean": 0.00274965228891233, "completions/clipped_ratio": 0.0, "completions/max_length": 586.9, "completions/max_terminated_length": 586.9, "completions/mean_length": 206.12187957763672, "completions/mean_terminated_length": 206.12187957763672, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.032186857688551146, "epoch": 21.07774716865626, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3524928097.0, "reward": 8.006720161437988, "reward_std": 0.29059007614850996, "rewards/combined_reward/mean": 8.006720066070557, "rewards/combined_reward/std": 2.317874586582184, "step": 34440 }, { "clip_ratio/high_max": 0.001983882985950913, "clip_ratio/high_mean": 0.0011290658770424973, "clip_ratio/low_mean": 0.0009250090279238066, "clip_ratio/low_min": 0.00023055164007625232, "clip_ratio/region_mean": 0.0020540748810162768, "completions/clipped_ratio": 0.0, "completions/max_length": 545.5, "completions/max_terminated_length": 545.5, "completions/mean_length": 196.3843780517578, "completions/mean_terminated_length": 196.3843780517578, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.02670520539395511, "epoch": 21.09611264156719, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3527982373.0, "reward": 8.099881839752197, "reward_std": 0.33859655559062957, "rewards/combined_reward/mean": 8.099881505966186, "rewards/combined_reward/std": 2.34927761554718, "step": 34470 }, { "clip_ratio/high_max": 0.0021391163347288964, "clip_ratio/high_mean": 0.0012776118635277574, "clip_ratio/low_mean": 0.0010198922752654956, "clip_ratio/low_min": 0.00027287408180806476, "clip_ratio/region_mean": 0.002297504181721403, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 200.1867218017578, "completions/mean_terminated_length": 200.1867218017578, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.02891808347776532, "epoch": 21.114478114478114, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3531051250.0, "reward": 8.232715082168578, "reward_std": 0.32023500502109525, "rewards/combined_reward/mean": 8.23271484375, "rewards/combined_reward/std": 2.0848337411880493, "step": 34500 }, { "clip_ratio/high_max": 0.0018501888383373928, "clip_ratio/high_mean": 0.0011275694739500371, "clip_ratio/low_mean": 0.001134339145392005, "clip_ratio/low_min": 0.00037679163797292857, "clip_ratio/region_mean": 0.002261908613824441, "completions/clipped_ratio": 0.0, "completions/max_length": 590.3, "completions/max_terminated_length": 590.3, "completions/mean_length": 214.67578582763673, "completions/mean_terminated_length": 214.67578582763673, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.026589213660918175, "epoch": 21.132843587389043, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3534175765.0, "reward": 8.36085810661316, "reward_std": 0.3144115224480629, "rewards/combined_reward/mean": 8.360857582092285, "rewards/combined_reward/std": 1.9210549235343932, "step": 34530 }, { "clip_ratio/high_max": 0.002359486229640121, "clip_ratio/high_mean": 0.001249953097430989, "clip_ratio/low_mean": 0.0009733003536287773, "clip_ratio/low_min": 0.00020646241415912906, "clip_ratio/region_mean": 0.0022232534470579897, "completions/clipped_ratio": 0.0, "completions/max_length": 686.5, "completions/max_terminated_length": 686.5, "completions/mean_length": 187.1893280029297, "completions/mean_terminated_length": 187.1893280029297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.031688754664113125, "epoch": 21.15120906029997, "frac_reward_zero_std": 0.18333333916962147, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3537194732.0, "reward": 7.987733840942383, "reward_std": 0.3096190720796585, "rewards/combined_reward/mean": 7.987733650207519, "rewards/combined_reward/std": 2.195101261138916, "step": 34560 }, { "clip_ratio/high_max": 0.0028455069962850152, "clip_ratio/high_mean": 0.001527483951698135, "clip_ratio/low_mean": 0.0010747201379369168, "clip_ratio/low_min": 0.0003352996738006671, "clip_ratio/region_mean": 0.0026022040867246686, "completions/clipped_ratio": 0.0, "completions/max_length": 739.3, "completions/max_terminated_length": 739.3, "completions/mean_length": 199.56849670410156, "completions/mean_terminated_length": 199.56849670410156, "completions/min_length": 39.2, "completions/min_terminated_length": 39.2, "entropy": 0.03164836800036331, "epoch": 21.1695745332109, "frac_reward_zero_std": 0.18750000223517418, "grad_norm": 0.0242919921875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3540261235.0, "reward": 7.891921138763427, "reward_std": 0.35235586017370224, "rewards/combined_reward/mean": 7.891921091079712, "rewards/combined_reward/std": 2.495885932445526, "step": 34590 }, { "clip_ratio/high_max": 0.0026862321072258057, "clip_ratio/high_mean": 0.001430484942466137, "clip_ratio/low_mean": 0.0009527195519088613, "clip_ratio/low_min": 0.00020681598883432646, "clip_ratio/region_mean": 0.002383204487826636, "completions/clipped_ratio": 0.0, "completions/max_length": 501.4, "completions/max_terminated_length": 501.4, "completions/mean_length": 186.30312957763672, "completions/mean_terminated_length": 186.30312957763672, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027396177329743903, "epoch": 21.187940006121824, "frac_reward_zero_std": 0.20416666977107525, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3543276799.0, "reward": 7.844000768661499, "reward_std": 0.345738060772419, "rewards/combined_reward/mean": 7.844000244140625, "rewards/combined_reward/std": 2.6857600927352907, "step": 34620 }, { "clip_ratio/high_max": 0.002324729919200763, "clip_ratio/high_mean": 0.0012634059826571804, "clip_ratio/low_mean": 0.001016886993481118, "clip_ratio/low_min": 0.0002613924288501342, "clip_ratio/region_mean": 0.0022802929840205857, "completions/clipped_ratio": 0.0, "completions/max_length": 602.4, "completions/max_terminated_length": 602.4, "completions/mean_length": 196.7947982788086, "completions/mean_terminated_length": 196.7947982788086, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.027320156913871567, "epoch": 21.206305479032753, "frac_reward_zero_std": 0.17083334177732468, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3546332651.0, "reward": 8.109923791885375, "reward_std": 0.3201366186141968, "rewards/combined_reward/mean": 8.109923362731934, "rewards/combined_reward/std": 2.1945948362350465, "step": 34650 }, { "clip_ratio/high_max": 0.003049709766249483, "clip_ratio/high_mean": 0.001663533163567384, "clip_ratio/low_mean": 0.0008996937016490847, "clip_ratio/low_min": 0.00019106551432438815, "clip_ratio/region_mean": 0.0025632268593957027, "completions/clipped_ratio": 0.0, "completions/max_length": 497.6, "completions/max_terminated_length": 497.6, "completions/mean_length": 200.86719207763673, "completions/mean_terminated_length": 200.86719207763673, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.029505980713292957, "epoch": 21.22467095194368, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 3549404141.0, "reward": 7.98542275428772, "reward_std": 0.2331373706459999, "rewards/combined_reward/mean": 7.985422515869141, "rewards/combined_reward/std": 2.374983048439026, "step": 34680 }, { "clip_ratio/high_max": 0.0025319480497273616, "clip_ratio/high_mean": 0.0014652840195291598, "clip_ratio/low_mean": 0.0012117597615239597, "clip_ratio/low_min": 0.0003428688888864902, "clip_ratio/region_mean": 0.0026770437844485665, "completions/clipped_ratio": 0.0, "completions/max_length": 589.9, "completions/max_terminated_length": 589.9, "completions/mean_length": 199.08412017822266, "completions/mean_terminated_length": 199.08412017822266, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.028944586698586742, "epoch": 21.24303642485461, "frac_reward_zero_std": 0.15833333767950536, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3552468784.0, "reward": 8.076677417755127, "reward_std": 0.3350858435034752, "rewards/combined_reward/mean": 8.076676988601685, "rewards/combined_reward/std": 2.250074803829193, "step": 34710 }, { "clip_ratio/high_max": 0.002304207173923108, "clip_ratio/high_mean": 0.0014933618826034945, "clip_ratio/low_mean": 0.0009579686906363349, "clip_ratio/low_min": 0.0003031020605703816, "clip_ratio/region_mean": 0.0024513305635385525, "completions/clipped_ratio": 0.0, "completions/max_length": 541.5, "completions/max_terminated_length": 541.5, "completions/mean_length": 197.61380767822266, "completions/mean_terminated_length": 197.61380767822266, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028721333481371403, "epoch": 21.261401897765534, "frac_reward_zero_std": 0.15833333767950536, "grad_norm": 0.0050048828125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3555527781.0, "reward": 8.298969793319703, "reward_std": 0.34496202394366265, "rewards/combined_reward/mean": 8.298969411849976, "rewards/combined_reward/std": 1.9040901064872742, "step": 34740 }, { "clip_ratio/high_max": 0.002619038569779756, "clip_ratio/high_mean": 0.0014984745879094892, "clip_ratio/low_mean": 0.0008695806653122417, "clip_ratio/low_min": 0.00026717122406504735, "clip_ratio/region_mean": 0.0023680552636506035, "completions/clipped_ratio": 0.0, "completions/max_length": 558.9, "completions/max_terminated_length": 558.9, "completions/mean_length": 200.2085983276367, "completions/mean_terminated_length": 200.2085983276367, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027886515017598867, "epoch": 21.27976737067646, "frac_reward_zero_std": 0.1416666705161333, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3558596742.0, "reward": 8.148392486572266, "reward_std": 0.34207051396369936, "rewards/combined_reward/mean": 8.14839220046997, "rewards/combined_reward/std": 2.218370246887207, "step": 34770 }, { "clip_ratio/high_max": 0.002171425618629049, "clip_ratio/high_mean": 0.0011581142180754492, "clip_ratio/low_mean": 0.000928939940968121, "clip_ratio/low_min": 0.0001977863782182491, "clip_ratio/region_mean": 0.00208705415619382, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 194.96224365234374, "completions/mean_terminated_length": 194.96224365234374, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.025035236418868106, "epoch": 21.29813284358739, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3561645557.0, "reward": 8.4731436252594, "reward_std": 0.28024310767650606, "rewards/combined_reward/mean": 8.473143243789673, "rewards/combined_reward/std": 1.7137880682945252, "step": 34800 }, { "clip_ratio/high_max": 0.0022981685823954953, "clip_ratio/high_mean": 0.0012913004489140197, "clip_ratio/low_mean": 0.001014847842695114, "clip_ratio/low_min": 0.000270957588024127, "clip_ratio/region_mean": 0.0023061483031293997, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 201.2296920776367, "completions/mean_terminated_length": 201.2296920776367, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.0318803753859053, "epoch": 21.316498316498315, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3564718439.0, "reward": 7.955046033859253, "reward_std": 0.31432574391365053, "rewards/combined_reward/mean": 7.955045747756958, "rewards/combined_reward/std": 2.3569414258003234, "step": 34830 }, { "clip_ratio/high_max": 0.0026291164928503954, "clip_ratio/high_mean": 0.0014304681884823368, "clip_ratio/low_mean": 0.001028237468805552, "clip_ratio/low_min": 0.00036064512169105, "clip_ratio/region_mean": 0.002458705662866123, "completions/clipped_ratio": 0.0, "completions/max_length": 531.4, "completions/max_terminated_length": 531.4, "completions/mean_length": 203.75625305175782, "completions/mean_terminated_length": 203.75625305175782, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.029670161893591285, "epoch": 21.334863789409244, "frac_reward_zero_std": 0.14166666865348815, "grad_norm": 0.00787353515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3567801023.0, "reward": 7.899670648574829, "reward_std": 0.2938836731016636, "rewards/combined_reward/mean": 7.899670600891113, "rewards/combined_reward/std": 2.523632788658142, "step": 34860 }, { "clip_ratio/high_max": 0.002402981461151891, "clip_ratio/high_mean": 0.0013455769604358162, "clip_ratio/low_mean": 0.0007623548915338082, "clip_ratio/low_min": 0.0001909878500252186, "clip_ratio/region_mean": 0.0021079318520302575, "completions/clipped_ratio": 0.0, "completions/max_length": 555.8, "completions/max_terminated_length": 555.8, "completions/mean_length": 198.3921920776367, "completions/mean_terminated_length": 198.3921920776367, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.02767899351504942, "epoch": 21.35322926232017, "frac_reward_zero_std": 0.1750000026077032, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3570863009.0, "reward": 8.074211025238037, "reward_std": 0.31326671987771987, "rewards/combined_reward/mean": 8.074210453033448, "rewards/combined_reward/std": 2.3089853644371034, "step": 34890 }, { "clip_ratio/high_max": 0.002806349717381333, "clip_ratio/high_mean": 0.0014888399859046332, "clip_ratio/low_mean": 0.0009990675213581806, "clip_ratio/low_min": 0.0002734046184438436, "clip_ratio/region_mean": 0.0024879074956212816, "completions/clipped_ratio": 0.0, "completions/max_length": 645.7, "completions/max_terminated_length": 645.7, "completions/mean_length": 200.02083892822264, "completions/mean_terminated_length": 200.02083892822264, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02999878991395235, "epoch": 21.3715947352311, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.004608154296875, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3573931249.0, "reward": 8.439788055419921, "reward_std": 0.3230362944304943, "rewards/combined_reward/mean": 8.439788007736206, "rewards/combined_reward/std": 1.7607338726520538, "step": 34920 }, { "clip_ratio/high_max": 0.0023092625522015925, "clip_ratio/high_mean": 0.001313013077985185, "clip_ratio/low_mean": 0.0010497215612607154, "clip_ratio/low_min": 0.0003491511051834095, "clip_ratio/region_mean": 0.0023627346274831024, "completions/clipped_ratio": 0.0, "completions/max_length": 588.7, "completions/max_terminated_length": 588.7, "completions/mean_length": 199.90469207763672, "completions/mean_terminated_length": 199.90469207763672, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.0270215218498682, "epoch": 21.389960208142025, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 3576999043.0, "reward": 7.888049030303955, "reward_std": 0.32713259309530257, "rewards/combined_reward/mean": 7.888048839569092, "rewards/combined_reward/std": 2.4367836833000185, "step": 34950 }, { "clip_ratio/high_max": 0.0026584275265728744, "clip_ratio/high_mean": 0.0014876924654042038, "clip_ratio/low_mean": 0.0009882816550089047, "clip_ratio/low_min": 0.0003488227108997914, "clip_ratio/region_mean": 0.00247597413447996, "completions/clipped_ratio": 0.0, "completions/max_length": 611.3, "completions/max_terminated_length": 611.3, "completions/mean_length": 201.52708892822267, "completions/mean_terminated_length": 201.52708892822267, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03147182702086866, "epoch": 21.408325681052954, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.025634765625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3580073067.0, "reward": 7.923998546600342, "reward_std": 0.32271908223629, "rewards/combined_reward/mean": 7.92399845123291, "rewards/combined_reward/std": 2.3119133710861206, "step": 34980 }, { "clip_ratio/high_max": 0.002280866604996845, "clip_ratio/high_mean": 0.001339069229531257, "clip_ratio/low_mean": 0.0009859168336333823, "clip_ratio/low_min": 0.0003159479805617593, "clip_ratio/region_mean": 0.00232498607171389, "completions/clipped_ratio": 0.0, "completions/max_length": 540.7, "completions/max_terminated_length": 540.7, "completions/mean_length": 198.97865295410156, "completions/mean_terminated_length": 198.97865295410156, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.026712512721618016, "epoch": 21.42669115396388, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3583137305.0, "reward": 8.179882383346557, "reward_std": 0.29611545354127883, "rewards/combined_reward/mean": 8.17988224029541, "rewards/combined_reward/std": 2.215551292896271, "step": 35010 }, { "clip_ratio/high_max": 0.0023321252917715657, "clip_ratio/high_mean": 0.0012746205127768917, "clip_ratio/low_mean": 0.0006941979668530014, "clip_ratio/low_min": 0.00018573403140180743, "clip_ratio/region_mean": 0.001968818495394468, "completions/clipped_ratio": 0.0, "completions/max_length": 575.8, "completions/max_terminated_length": 575.8, "completions/mean_length": 192.29271240234374, "completions/mean_terminated_length": 192.29271240234374, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.025095580642422042, "epoch": 21.44505662687481, "frac_reward_zero_std": 0.22500000521540642, "grad_norm": 0.018798828125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3586175869.0, "reward": 8.198025560379028, "reward_std": 0.29867408126592637, "rewards/combined_reward/mean": 8.198025035858155, "rewards/combined_reward/std": 2.2287442088127136, "step": 35040 }, { "clip_ratio/high_max": 0.002442290419033573, "clip_ratio/high_mean": 0.0015245360390205556, "clip_ratio/low_mean": 0.0008674941012941417, "clip_ratio/low_min": 0.00030489724546593305, "clip_ratio/region_mean": 0.002392030131886713, "completions/clipped_ratio": 0.0, "completions/max_length": 597.1, "completions/max_terminated_length": 597.1, "completions/mean_length": 201.7453170776367, "completions/mean_terminated_length": 201.7453170776367, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.0289377188620468, "epoch": 21.463422099785735, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 3589250731.0, "reward": 8.098867607116699, "reward_std": 0.34449884593486785, "rewards/combined_reward/mean": 8.098867225646973, "rewards/combined_reward/std": 2.348194348812103, "step": 35070 }, { "clip_ratio/high_max": 0.00240603060471282, "clip_ratio/high_mean": 0.0013848588976543396, "clip_ratio/low_mean": 0.0008651597661203899, "clip_ratio/low_min": 0.0002374076085591999, "clip_ratio/region_mean": 0.0022500186891799482, "completions/clipped_ratio": 0.0, "completions/max_length": 588.3, "completions/max_terminated_length": 588.3, "completions/mean_length": 202.06328582763672, "completions/mean_terminated_length": 202.06328582763672, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.027801143059817453, "epoch": 21.481787572696664, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3592326814.0, "reward": 8.10695776939392, "reward_std": 0.2986296139657497, "rewards/combined_reward/mean": 8.10695767402649, "rewards/combined_reward/std": 2.1421167850494385, "step": 35100 }, { "clip_ratio/high_max": 0.002134298959814866, "clip_ratio/high_mean": 0.0012323684228855807, "clip_ratio/low_mean": 0.0009871065171561592, "clip_ratio/low_min": 0.00031372143882132757, "clip_ratio/region_mean": 0.0022194749554425167, "completions/clipped_ratio": 0.0, "completions/max_length": 510.5, "completions/max_terminated_length": 510.5, "completions/mean_length": 195.25989990234376, "completions/mean_terminated_length": 195.25989990234376, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "entropy": 0.025604343918773035, "epoch": 21.50015304560759, "frac_reward_zero_std": 0.2333333410322666, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3595376772.0, "reward": 7.988764381408691, "reward_std": 0.33678188920021057, "rewards/combined_reward/mean": 7.9887641906738285, "rewards/combined_reward/std": 2.439428114891052, "step": 35130 }, { "clip_ratio/high_max": 0.0018535270489034399, "clip_ratio/high_mean": 0.0010565877083233014, "clip_ratio/low_mean": 0.0008500283906566135, "clip_ratio/low_min": 0.00021414477693421457, "clip_ratio/region_mean": 0.0019066160937654785, "completions/clipped_ratio": 0.0, "completions/max_length": 526.3, "completions/max_terminated_length": 526.3, "completions/mean_length": 204.51276702880858, "completions/mean_terminated_length": 204.51276702880858, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.02503205242101103, "epoch": 21.51851851851852, "frac_reward_zero_std": 0.20833333656191827, "grad_norm": 0.0252685546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3598462261.0, "reward": 8.06783323287964, "reward_std": 0.32580071687698364, "rewards/combined_reward/mean": 8.067832899093627, "rewards/combined_reward/std": 2.116621744632721, "step": 35160 }, { "clip_ratio/high_max": 0.0023714979019132444, "clip_ratio/high_mean": 0.0014449435662148365, "clip_ratio/low_mean": 0.0009241347306669923, "clip_ratio/low_min": 0.0003033894509523331, "clip_ratio/region_mean": 0.0023690782982157543, "completions/clipped_ratio": 0.0, "completions/max_length": 575.4, "completions/max_terminated_length": 575.4, "completions/mean_length": 204.41901397705078, "completions/mean_terminated_length": 204.41901397705078, "completions/min_length": 47.7, "completions/min_terminated_length": 47.7, "entropy": 0.029507925966754556, "epoch": 21.536883991429445, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.005401611328125, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3601547390.0, "reward": 8.078850936889648, "reward_std": 0.26850111186504366, "rewards/combined_reward/mean": 8.07885046005249, "rewards/combined_reward/std": 2.4073730587959288, "step": 35190 }, { "clip_ratio/high_max": 0.002903289390572657, "clip_ratio/high_mean": 0.001680667720453736, "clip_ratio/low_mean": 0.0009480212490113141, "clip_ratio/low_min": 0.0002670358204341028, "clip_ratio/region_mean": 0.002628688951275156, "completions/clipped_ratio": 0.0, "completions/max_length": 589.7, "completions/max_terminated_length": 589.7, "completions/mean_length": 195.37865142822267, "completions/mean_terminated_length": 195.37865142822267, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030517221614718437, "epoch": 21.555249464340374, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 3604597804.0, "reward": 8.023205661773682, "reward_std": 0.293517030775547, "rewards/combined_reward/mean": 8.023205327987672, "rewards/combined_reward/std": 2.3929550528526304, "step": 35220 }, { "clip_ratio/high_max": 0.0023284656926989555, "clip_ratio/high_mean": 0.0013201250056833183, "clip_ratio/low_mean": 0.0010779369073134148, "clip_ratio/low_min": 0.00032911091402638705, "clip_ratio/region_mean": 0.0023980619153007865, "completions/clipped_ratio": 0.0, "completions/max_length": 534.1, "completions/max_terminated_length": 534.1, "completions/mean_length": 191.8960983276367, "completions/mean_terminated_length": 191.8960983276367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028485180484130977, "epoch": 21.5736149372513, "frac_reward_zero_std": 0.20833334252238273, "grad_norm": 0.006134033203125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3607634845.0, "reward": 8.091394710540772, "reward_std": 0.269575434923172, "rewards/combined_reward/mean": 8.091394519805908, "rewards/combined_reward/std": 2.285622775554657, "step": 35250 }, { "clip_ratio/high_max": 0.0023346790791644403, "clip_ratio/high_mean": 0.0013885355975313965, "clip_ratio/low_mean": 0.0010238148356923678, "clip_ratio/low_min": 0.00021739942603744568, "clip_ratio/region_mean": 0.00241235043455769, "completions/clipped_ratio": 0.0, "completions/max_length": 579.7, "completions/max_terminated_length": 579.7, "completions/mean_length": 192.08802795410156, "completions/mean_terminated_length": 192.08802795410156, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02946508568711579, "epoch": 21.59198041016223, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3610672623.0, "reward": 7.9297645568847654, "reward_std": 0.3380042791366577, "rewards/combined_reward/mean": 7.929764032363892, "rewards/combined_reward/std": 2.452813374996185, "step": 35280 }, { "clip_ratio/high_max": 0.003050047543850572, "clip_ratio/high_mean": 0.0016744152286264579, "clip_ratio/low_mean": 0.0009123737870443923, "clip_ratio/low_min": 0.0001787774451562048, "clip_ratio/region_mean": 0.0025867890149432545, "completions/clipped_ratio": 0.0, "completions/max_length": 591.8, "completions/max_terminated_length": 591.8, "completions/mean_length": 200.58255615234376, "completions/mean_terminated_length": 200.58255615234376, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.032236264149347944, "epoch": 21.610345883073155, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 3613743020.0, "reward": 8.024079704284668, "reward_std": 0.36060353368520737, "rewards/combined_reward/mean": 8.024079751968383, "rewards/combined_reward/std": 2.1398055613040925, "step": 35310 }, { "clip_ratio/high_max": 0.0027118425433097096, "clip_ratio/high_mean": 0.0016039252363649818, "clip_ratio/low_mean": 0.0010860171606812705, "clip_ratio/low_min": 0.00036386254214448854, "clip_ratio/region_mean": 0.0026899424294242634, "completions/clipped_ratio": 0.0, "completions/max_length": 573.8, "completions/max_terminated_length": 573.8, "completions/mean_length": 198.3546920776367, "completions/mean_terminated_length": 198.3546920776367, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.03359721276598672, "epoch": 21.628711355984084, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 3616804862.0, "reward": 7.996551179885865, "reward_std": 0.3945869833230972, "rewards/combined_reward/mean": 7.996550989151001, "rewards/combined_reward/std": 2.2863780856132507, "step": 35340 }, { "clip_ratio/high_max": 0.0026212755207476826, "clip_ratio/high_mean": 0.0015136092518029424, "clip_ratio/low_mean": 0.001045823289071753, "clip_ratio/low_min": 0.00024630034046519236, "clip_ratio/region_mean": 0.002559432557124334, "completions/clipped_ratio": 0.0, "completions/max_length": 599.2, "completions/max_terminated_length": 599.2, "completions/mean_length": 197.44609832763672, "completions/mean_terminated_length": 197.44609832763672, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.032072755383948486, "epoch": 21.64707682889501, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.00726318359375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3619863215.0, "reward": 8.26178469657898, "reward_std": 0.2376369059085846, "rewards/combined_reward/mean": 8.261784315109253, "rewards/combined_reward/std": 2.1001747727394102, "step": 35370 }, { "clip_ratio/high_max": 0.002730714072337529, "clip_ratio/high_mean": 0.0014855337438348215, "clip_ratio/low_mean": 0.0010366648023894716, "clip_ratio/low_min": 0.00032601211860310285, "clip_ratio/region_mean": 0.0025221985633834265, "completions/clipped_ratio": 0.0, "completions/max_length": 606.1, "completions/max_terminated_length": 606.1, "completions/mean_length": 198.10833740234375, "completions/mean_terminated_length": 198.10833740234375, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03288292145977418, "epoch": 21.66544230180594, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.01416015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3622924111.0, "reward": 7.790680456161499, "reward_std": 0.4048777222633362, "rewards/combined_reward/mean": 7.79068021774292, "rewards/combined_reward/std": 2.3653801918029784, "step": 35400 }, { "clip_ratio/high_max": 0.00283986299133782, "clip_ratio/high_mean": 0.0014861859976614748, "clip_ratio/low_mean": 0.0009001165992231109, "clip_ratio/low_min": 0.00019983404915061934, "clip_ratio/region_mean": 0.0023863026033116813, "completions/clipped_ratio": 0.0, "completions/max_length": 569.8, "completions/max_terminated_length": 569.8, "completions/mean_length": 198.06537017822265, "completions/mean_terminated_length": 198.06537017822265, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.02946240163097779, "epoch": 21.683807774716865, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3625984842.0, "reward": 8.160041809082031, "reward_std": 0.3064802512526512, "rewards/combined_reward/mean": 8.16004147529602, "rewards/combined_reward/std": 2.1426028490066527, "step": 35430 }, { "clip_ratio/high_max": 0.001992765577839843, "clip_ratio/high_mean": 0.0010477639758998218, "clip_ratio/low_mean": 0.000971371864216053, "clip_ratio/low_min": 0.0002583991651287458, "clip_ratio/region_mean": 0.0020191358564867795, "completions/clipped_ratio": 0.0, "completions/max_length": 537.9, "completions/max_terminated_length": 537.9, "completions/mean_length": 193.40677490234376, "completions/mean_terminated_length": 193.40677490234376, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027999461535364388, "epoch": 21.702173247627794, "frac_reward_zero_std": 0.14583333507180213, "grad_norm": 0.0159912109375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3629027684.0, "reward": 8.035334539413451, "reward_std": 0.28527589440345763, "rewards/combined_reward/mean": 8.035334205627441, "rewards/combined_reward/std": 2.3507869839668274, "step": 35460 }, { "clip_ratio/high_max": 0.002518944563053083, "clip_ratio/high_mean": 0.0015070953702282471, "clip_ratio/low_mean": 0.0009837492341224182, "clip_ratio/low_min": 0.0003346512634986235, "clip_ratio/region_mean": 0.0024908446105352294, "completions/clipped_ratio": 0.0, "completions/max_length": 586.9, "completions/max_terminated_length": 586.9, "completions/mean_length": 200.71068115234374, "completions/mean_terminated_length": 200.71068115234374, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.031294207011039056, "epoch": 21.72053872053872, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3632098573.0, "reward": 7.990128850936889, "reward_std": 0.3138268902897835, "rewards/combined_reward/mean": 7.99012861251831, "rewards/combined_reward/std": 2.2924099922180177, "step": 35490 }, { "clip_ratio/high_max": 0.0028114295069826766, "clip_ratio/high_mean": 0.0015909083313696707, "clip_ratio/low_mean": 0.0010123477071222926, "clip_ratio/low_min": 0.0002838395092112478, "clip_ratio/region_mean": 0.002603256030124612, "completions/clipped_ratio": 0.0, "completions/max_length": 628.8, "completions/max_terminated_length": 628.8, "completions/mean_length": 198.14349517822265, "completions/mean_terminated_length": 198.14349517822265, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03221152110490948, "epoch": 21.73890419344965, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0027, "num_tokens": 3635159604.0, "reward": 7.764494562149048, "reward_std": 0.3638666749000549, "rewards/combined_reward/mean": 7.764494180679321, "rewards/combined_reward/std": 2.474236857891083, "step": 35520 }, { "clip_ratio/high_max": 0.0032065548890386707, "clip_ratio/high_mean": 0.001888035198013919, "clip_ratio/low_mean": 0.0009681542593170889, "clip_ratio/low_min": 0.0002709730976979093, "clip_ratio/region_mean": 0.002856189456845944, "completions/clipped_ratio": 0.0, "completions/max_length": 576.9, "completions/max_terminated_length": 576.9, "completions/mean_length": 203.9018295288086, "completions/mean_terminated_length": 203.9018295288086, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.033332955092191696, "epoch": 21.757269666360575, "frac_reward_zero_std": 0.18750000298023223, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3638242747.0, "reward": 7.9876504898071286, "reward_std": 0.3132962487637997, "rewards/combined_reward/mean": 7.987650346755982, "rewards/combined_reward/std": 2.3475972175598145, "step": 35550 }, { "clip_ratio/high_max": 0.0022440036467742176, "clip_ratio/high_mean": 0.0013514063156132275, "clip_ratio/low_mean": 0.0009619777334592073, "clip_ratio/low_min": 0.0002664658102730755, "clip_ratio/region_mean": 0.0023133840603501692, "completions/clipped_ratio": 0.0, "completions/max_length": 661.5, "completions/max_terminated_length": 661.5, "completions/mean_length": 196.17630310058593, "completions/mean_terminated_length": 196.17630310058593, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028590585539738336, "epoch": 21.775635139271504, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3641296224.0, "reward": 8.119771814346313, "reward_std": 0.3268084332346916, "rewards/combined_reward/mean": 8.11977152824402, "rewards/combined_reward/std": 2.230665934085846, "step": 35580 }, { "clip_ratio/high_max": 0.0020633398613426833, "clip_ratio/high_mean": 0.0009306558153184596, "clip_ratio/low_mean": 0.00035395747818256496, "clip_ratio/low_min": 4.2724207014543934e-05, "clip_ratio/region_mean": 0.001284613298776094, "completions/clipped_ratio": 0.0, "completions/max_length": 525.25, "completions/max_terminated_length": 525.25, "completions/mean_length": 203.13802337646484, "completions/mean_terminated_length": 203.13802337646484, "completions/min_length": 37.25, "completions/min_terminated_length": 37.25, "entropy": 0.025364818517118692, "epoch": 21.793388429752067, "frac_reward_zero_std": 0.1458333395421505, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3644384128.0, "reward": 7.960522532463074, "reward_std": 0.3184201568365097, "rewards/combined_reward/mean": 7.960522532463074, "rewards/combined_reward/std": 2.5318875312805176, "step": 35610 }, { "clip_ratio/high_max": 0.0016738265156163834, "clip_ratio/high_mean": 0.0009215775030800917, "clip_ratio/low_mean": 0.000867633076995844, "clip_ratio/low_min": 0.00027680482332167837, "clip_ratio/region_mean": 0.0017892105689194674, "completions/clipped_ratio": 0.0, "completions/max_length": 629.8, "completions/max_terminated_length": 629.8, "completions/mean_length": 204.57188262939454, "completions/mean_terminated_length": 204.57188262939454, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.027658209235717855, "epoch": 21.811753902662993, "frac_reward_zero_std": 0.17500000707805158, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3647469844.0, "reward": 8.33563289642334, "reward_std": 0.3067042797803879, "rewards/combined_reward/mean": 8.335632562637329, "rewards/combined_reward/std": 1.9107816576957704, "step": 35640 }, { "clip_ratio/high_max": 0.0025295230276242363, "clip_ratio/high_mean": 0.0014247770355723332, "clip_ratio/low_mean": 0.0008185024556344919, "clip_ratio/low_min": 0.00021454301677295006, "clip_ratio/region_mean": 0.002243279487690112, "completions/clipped_ratio": 0.0, "completions/max_length": 529.2, "completions/max_terminated_length": 529.2, "completions/mean_length": 194.70937957763672, "completions/mean_terminated_length": 194.70937957763672, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.03278294044236342, "epoch": 21.830119375573922, "frac_reward_zero_std": 0.2083333373069763, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3650517688.0, "reward": 8.027589702606202, "reward_std": 0.30641250908374784, "rewards/combined_reward/mean": 8.027589321136475, "rewards/combined_reward/std": 2.242492914199829, "step": 35670 }, { "clip_ratio/high_max": 0.001672704841378921, "clip_ratio/high_mean": 0.0009675598189157124, "clip_ratio/low_mean": 0.0007051575254687729, "clip_ratio/low_min": 0.00024039716202726898, "clip_ratio/region_mean": 0.0016727173385637192, "completions/clipped_ratio": 0.0, "completions/max_length": 726.2, "completions/max_terminated_length": 726.2, "completions/mean_length": 193.61823577880858, "completions/mean_terminated_length": 193.61823577880858, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.025389285750376682, "epoch": 21.848484848484848, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3653561342.0, "reward": 8.143060541152954, "reward_std": 0.31597366333007815, "rewards/combined_reward/mean": 8.143060445785522, "rewards/combined_reward/std": 2.2062344670295717, "step": 35700 }, { "clip_ratio/high_max": 0.0018402337125735357, "clip_ratio/high_mean": 0.0011251282803035187, "clip_ratio/low_mean": 0.000707751124718925, "clip_ratio/low_min": 0.00024648975571229436, "clip_ratio/region_mean": 0.0018328794074477627, "completions/clipped_ratio": 0.0, "completions/max_length": 549.5, "completions/max_terminated_length": 549.5, "completions/mean_length": 201.9041717529297, "completions/mean_terminated_length": 201.9041717529297, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.030699247602994244, "epoch": 21.866850321395777, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3656636814.0, "reward": 8.351690101623536, "reward_std": 0.35622200816869737, "rewards/combined_reward/mean": 8.351690006256103, "rewards/combined_reward/std": 1.945130717754364, "step": 35730 }, { "clip_ratio/high_max": 0.0020913674719243622, "clip_ratio/high_mean": 0.0012168269751176317, "clip_ratio/low_mean": 0.0007215961399197113, "clip_ratio/low_min": 0.00018094226931377003, "clip_ratio/region_mean": 0.0019384231224345665, "completions/clipped_ratio": 0.0, "completions/max_length": 543.8, "completions/max_terminated_length": 543.8, "completions/mean_length": 188.30990142822264, "completions/mean_terminated_length": 188.30990142822264, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.03199436290500065, "epoch": 21.885215794306703, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3659660084.0, "reward": 8.04921488761902, "reward_std": 0.2907188802957535, "rewards/combined_reward/mean": 8.049214696884155, "rewards/combined_reward/std": 2.279698944091797, "step": 35760 }, { "clip_ratio/high_max": 0.0020222838284098543, "clip_ratio/high_mean": 0.001166744951478904, "clip_ratio/low_mean": 0.0007347890887710189, "clip_ratio/low_min": 0.00023355314745761765, "clip_ratio/region_mean": 0.0019015340375820719, "completions/clipped_ratio": 0.0, "completions/max_length": 595.2, "completions/max_terminated_length": 595.2, "completions/mean_length": 207.08203582763673, "completions/mean_terminated_length": 207.08203582763673, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.029359678373051187, "epoch": 21.903581267217632, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3662755439.0, "reward": 8.189321041107178, "reward_std": 0.3921765238046646, "rewards/combined_reward/mean": 8.18932065963745, "rewards/combined_reward/std": 2.1077464699745176, "step": 35790 }, { "clip_ratio/high_max": 0.0017577863298356533, "clip_ratio/high_mean": 0.0010821963055301846, "clip_ratio/low_mean": 0.0007462149779409326, "clip_ratio/low_min": 0.00022249865966538588, "clip_ratio/region_mean": 0.001828411271465787, "completions/clipped_ratio": 0.0, "completions/max_length": 550.7, "completions/max_terminated_length": 550.7, "completions/mean_length": 200.33698425292968, "completions/mean_terminated_length": 200.33698425292968, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.03032622141763568, "epoch": 21.921946740128558, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3665824893.0, "reward": 7.98438835144043, "reward_std": 0.33030247539281843, "rewards/combined_reward/mean": 7.984388208389282, "rewards/combined_reward/std": 2.449834907054901, "step": 35820 }, { "clip_ratio/high_max": 0.0018125297390118552, "clip_ratio/high_mean": 0.0010920992552807243, "clip_ratio/low_mean": 0.0007145034345739987, "clip_ratio/low_min": 0.0002503404718784926, "clip_ratio/region_mean": 0.0018066026978582764, "completions/clipped_ratio": 0.0, "completions/max_length": 558.2, "completions/max_terminated_length": 558.2, "completions/mean_length": 182.74219512939453, "completions/mean_terminated_length": 182.74219512939453, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02738372185267508, "epoch": 21.940312213039487, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3668826783.0, "reward": 7.793329334259033, "reward_std": 0.329296238720417, "rewards/combined_reward/mean": 7.793328905105591, "rewards/combined_reward/std": 2.4579827427864074, "step": 35850 }, { "clip_ratio/high_max": 0.0018952954371343367, "clip_ratio/high_mean": 0.001070468394633887, "clip_ratio/low_mean": 0.0007443311137952453, "clip_ratio/low_min": 0.00020160246688950186, "clip_ratio/region_mean": 0.0018147995108544516, "completions/clipped_ratio": 0.0, "completions/max_length": 562.5, "completions/max_terminated_length": 562.5, "completions/mean_length": 200.92969360351563, "completions/mean_terminated_length": 200.92969360351563, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02714264243841171, "epoch": 21.958677685950413, "frac_reward_zero_std": 0.12083333693444728, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3671898513.0, "reward": 8.07891125679016, "reward_std": 0.36447024941444395, "rewards/combined_reward/mean": 8.07891116142273, "rewards/combined_reward/std": 2.0889103293418883, "step": 35880 }, { "clip_ratio/high_max": 0.0017907488358711514, "clip_ratio/high_mean": 0.0009995488481460294, "clip_ratio/low_mean": 0.0007416058193484787, "clip_ratio/low_min": 0.00020388890601073703, "clip_ratio/region_mean": 0.0017411546825314872, "completions/clipped_ratio": 0.0, "completions/max_length": 515.7, "completions/max_terminated_length": 515.7, "completions/mean_length": 194.2830810546875, "completions/mean_terminated_length": 194.2830810546875, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.02893176389237245, "epoch": 21.977043158861342, "frac_reward_zero_std": 0.19166667014360428, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3674944720.0, "reward": 7.828337812423706, "reward_std": 0.30484976172447203, "rewards/combined_reward/mean": 7.828337574005127, "rewards/combined_reward/std": 2.5300180315971375, "step": 35910 }, { "clip_ratio/high_max": 0.002137530967593193, "clip_ratio/high_mean": 0.0012809721964003983, "clip_ratio/low_mean": 0.0008407680864668995, "clip_ratio/low_min": 0.00020622264370710278, "clip_ratio/region_mean": 0.0021217402982680747, "completions/clipped_ratio": 0.0, "completions/max_length": 644.6, "completions/max_terminated_length": 644.6, "completions/mean_length": 207.31250457763673, "completions/mean_terminated_length": 207.31250457763673, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.030098646615321436, "epoch": 21.995408631772268, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3678040960.0, "reward": 7.9619911193847654, "reward_std": 0.3340494453907013, "rewards/combined_reward/mean": 7.961991024017334, "rewards/combined_reward/std": 2.404939877986908, "step": 35940 }, { "clip_ratio/high_max": 0.0025310747336871066, "clip_ratio/high_mean": 0.001385007217501626, "clip_ratio/low_mean": 0.0008526898350506509, "clip_ratio/low_min": 0.00020708882232715726, "clip_ratio/region_mean": 0.0022376970589336495, "completions/clipped_ratio": 0.0, "completions/max_length": 610.7, "completions/max_terminated_length": 610.7, "completions/mean_length": 201.48802642822267, "completions/mean_terminated_length": 201.48802642822267, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029349022529652862, "epoch": 22.014080195898377, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3681114834.0, "reward": 8.135841035842896, "reward_std": 0.28343563824892043, "rewards/combined_reward/mean": 8.135840892791748, "rewards/combined_reward/std": 2.0217817783355714, "step": 35970 }, { "clip_ratio/high_max": 0.0024458921669671934, "clip_ratio/high_mean": 0.001354452909193545, "clip_ratio/low_mean": 0.0009446149682238077, "clip_ratio/low_min": 0.00028643539068677154, "clip_ratio/region_mean": 0.0022990678820254593, "completions/clipped_ratio": 0.0, "completions/max_length": 521.3, "completions/max_terminated_length": 521.3, "completions/mean_length": 203.18959045410156, "completions/mean_terminated_length": 203.18959045410156, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.029308298292259376, "epoch": 22.032445668809306, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3684195242.0, "reward": 7.917944145202637, "reward_std": 0.24690246284008027, "rewards/combined_reward/mean": 7.917943859100342, "rewards/combined_reward/std": 2.506607401371002, "step": 36000 }, { "clip_ratio/high_max": 0.002741040107503068, "clip_ratio/high_mean": 0.0015514250944155114, "clip_ratio/low_mean": 0.0010648569727588135, "clip_ratio/low_min": 0.00034231835622146415, "clip_ratio/region_mean": 0.0026162820693571122, "completions/clipped_ratio": 0.0, "completions/max_length": 659.2, "completions/max_terminated_length": 659.2, "completions/mean_length": 206.7804748535156, "completions/mean_terminated_length": 206.7804748535156, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03008752429547409, "epoch": 22.05081114172023, "frac_reward_zero_std": 0.1750000063329935, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3687289439.0, "reward": 8.110978412628175, "reward_std": 0.3084545716643333, "rewards/combined_reward/mean": 8.110977840423583, "rewards/combined_reward/std": 2.2866259932518007, "step": 36030 }, { "clip_ratio/high_max": 0.0026839241123525425, "clip_ratio/high_mean": 0.0015694133976163964, "clip_ratio/low_mean": 0.001048096862238405, "clip_ratio/low_min": 0.00034754360870768627, "clip_ratio/region_mean": 0.0026175102359654073, "completions/clipped_ratio": 0.0, "completions/max_length": 662.4, "completions/max_terminated_length": 662.4, "completions/mean_length": 205.77005920410156, "completions/mean_terminated_length": 205.77005920410156, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.03199329525232315, "epoch": 22.06917661463116, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3690379756.0, "reward": 8.043898248672486, "reward_std": 0.3023043319582939, "rewards/combined_reward/mean": 8.043898057937621, "rewards/combined_reward/std": 2.2180445551872254, "step": 36060 }, { "clip_ratio/high_max": 0.0027014940094280365, "clip_ratio/high_mean": 0.0015199359123168202, "clip_ratio/low_mean": 0.0010098980563877072, "clip_ratio/low_min": 0.0002832649729195206, "clip_ratio/region_mean": 0.0025298339596095806, "completions/clipped_ratio": 0.0, "completions/max_length": 557.6, "completions/max_terminated_length": 557.6, "completions/mean_length": 202.2205795288086, "completions/mean_terminated_length": 202.2205795288086, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029754042718559505, "epoch": 22.087542087542086, "frac_reward_zero_std": 0.1958333358168602, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3693456443.0, "reward": 8.10768165588379, "reward_std": 0.27567234337329866, "rewards/combined_reward/mean": 8.107681560516358, "rewards/combined_reward/std": 2.3307778596878053, "step": 36090 }, { "clip_ratio/high_max": 0.0017894057561837447, "clip_ratio/high_mean": 0.0009949782556456436, "clip_ratio/low_mean": 0.0009224664092471358, "clip_ratio/low_min": 0.00021348906763402436, "clip_ratio/region_mean": 0.001917444665741641, "completions/clipped_ratio": 0.0, "completions/max_length": 536.7, "completions/max_terminated_length": 536.7, "completions/mean_length": 196.93463897705078, "completions/mean_terminated_length": 196.93463897705078, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.02439749464392662, "epoch": 22.105907560453016, "frac_reward_zero_std": 0.1541666727513075, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3696512832.0, "reward": 8.073503541946412, "reward_std": 0.3679845333099365, "rewards/combined_reward/mean": 8.073503303527833, "rewards/combined_reward/std": 2.23700407743454, "step": 36120 }, { "clip_ratio/high_max": 0.0020138001690308254, "clip_ratio/high_mean": 0.001106245638341837, "clip_ratio/low_mean": 0.0009435171459093301, "clip_ratio/low_min": 0.00030648258471046577, "clip_ratio/region_mean": 0.0020497627915271248, "completions/clipped_ratio": 0.0, "completions/max_length": 527.2, "completions/max_terminated_length": 527.2, "completions/mean_length": 197.23099517822266, "completions/mean_terminated_length": 197.23099517822266, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.025928172903756302, "epoch": 22.12427303336394, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.004638671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3699570359.0, "reward": 8.282282972335816, "reward_std": 0.2636772558093071, "rewards/combined_reward/mean": 8.282283067703247, "rewards/combined_reward/std": 2.0726145267486573, "step": 36150 }, { "clip_ratio/high_max": 0.0022695385831563424, "clip_ratio/high_mean": 0.0013084900067042327, "clip_ratio/low_mean": 0.0009508507211042646, "clip_ratio/low_min": 0.00031684522982686756, "clip_ratio/region_mean": 0.0022593407396925614, "completions/clipped_ratio": 0.0, "completions/max_length": 684.1, "completions/max_terminated_length": 684.1, "completions/mean_length": 203.60365295410156, "completions/mean_terminated_length": 203.60365295410156, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.030205586552619935, "epoch": 22.14263850627487, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3702652357.0, "reward": 8.205664443969727, "reward_std": 0.33765062838792803, "rewards/combined_reward/mean": 8.2056640625, "rewards/combined_reward/std": 2.1169445753097533, "step": 36180 }, { "clip_ratio/high_max": 0.0028557423516758717, "clip_ratio/high_mean": 0.0016352741894176384, "clip_ratio/low_mean": 0.0009178410594055701, "clip_ratio/low_min": 0.00023279636491982576, "clip_ratio/region_mean": 0.0025531152646484164, "completions/clipped_ratio": 0.0, "completions/max_length": 646.2, "completions/max_terminated_length": 646.2, "completions/mean_length": 201.34219207763672, "completions/mean_terminated_length": 201.34219207763672, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03271858214090268, "epoch": 22.161003979185796, "frac_reward_zero_std": 0.18333333916962147, "grad_norm": 0.01068115234375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 3705725671.0, "reward": 8.047277450561523, "reward_std": 0.3344975173473358, "rewards/combined_reward/mean": 8.047277164459228, "rewards/combined_reward/std": 2.1505736470222474, "step": 36210 }, { "clip_ratio/high_max": 0.0025980577479155424, "clip_ratio/high_mean": 0.0013713334591254048, "clip_ratio/low_mean": 0.0010845303631867864, "clip_ratio/low_min": 0.00033649075267021543, "clip_ratio/region_mean": 0.002455863838258665, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 185.42187957763673, "completions/mean_terminated_length": 185.42187957763673, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.031480143591761586, "epoch": 22.179369452096726, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3708737851.0, "reward": 7.633085775375366, "reward_std": 0.3075350522994995, "rewards/combined_reward/mean": 7.633085489273071, "rewards/combined_reward/std": 2.8034391522407534, "step": 36240 }, { "clip_ratio/high_max": 0.002443715340147416, "clip_ratio/high_mean": 0.0013386859730720364, "clip_ratio/low_mean": 0.000757995489645206, "clip_ratio/low_min": 0.00019061912826146, "clip_ratio/region_mean": 0.002096681457866604, "completions/clipped_ratio": 0.0, "completions/max_length": 582.3, "completions/max_terminated_length": 582.3, "completions/mean_length": 198.07318267822265, "completions/mean_terminated_length": 198.07318267822265, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.02558202133513987, "epoch": 22.19773492500765, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3711798612.0, "reward": 8.0486328125, "reward_std": 0.3303725332021713, "rewards/combined_reward/mean": 8.048632669448853, "rewards/combined_reward/std": 2.451396870613098, "step": 36270 }, { "clip_ratio/high_max": 0.0025048215514592207, "clip_ratio/high_mean": 0.0014414109507924877, "clip_ratio/low_mean": 0.000965418642105457, "clip_ratio/low_min": 0.00022832075774203986, "clip_ratio/region_mean": 0.002406829582954136, "completions/clipped_ratio": 0.0, "completions/max_length": 524.6, "completions/max_terminated_length": 524.6, "completions/mean_length": 200.03307647705077, "completions/mean_terminated_length": 200.03307647705077, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029444329688946405, "epoch": 22.21610039791858, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3714866899.0, "reward": 8.154574823379516, "reward_std": 0.3047001212835312, "rewards/combined_reward/mean": 8.154574489593506, "rewards/combined_reward/std": 2.146818256378174, "step": 36300 }, { "clip_ratio/high_max": 0.0025667469245187627, "clip_ratio/high_mean": 0.0015308344993779125, "clip_ratio/low_mean": 0.0008375985652188926, "clip_ratio/low_min": 0.00018644086206525873, "clip_ratio/region_mean": 0.00236843307405555, "completions/clipped_ratio": 0.0, "completions/max_length": 572.3, "completions/max_terminated_length": 572.3, "completions/mean_length": 197.6968795776367, "completions/mean_terminated_length": 197.6968795776367, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03002945980988443, "epoch": 22.234465870829506, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.007080078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3717926215.0, "reward": 7.99808177947998, "reward_std": 0.31664976477622986, "rewards/combined_reward/mean": 7.99808144569397, "rewards/combined_reward/std": 2.3402413845062258, "step": 36330 }, { "clip_ratio/high_max": 0.0019544811448819624, "clip_ratio/high_mean": 0.001182761066229432, "clip_ratio/low_mean": 0.0009999522288126172, "clip_ratio/low_min": 0.000279391543881502, "clip_ratio/region_mean": 0.002182713303773198, "completions/clipped_ratio": 0.0, "completions/max_length": 588.9, "completions/max_terminated_length": 588.9, "completions/mean_length": 196.5031311035156, "completions/mean_terminated_length": 196.5031311035156, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.02816025187106182, "epoch": 22.252831343740436, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 3720980947.0, "reward": 8.231179809570312, "reward_std": 0.2817422941327095, "rewards/combined_reward/mean": 8.231179523468018, "rewards/combined_reward/std": 2.0649896621704102, "step": 36360 }, { "clip_ratio/high_max": 0.0021566093567040906, "clip_ratio/high_mean": 0.001369088634479946, "clip_ratio/low_mean": 0.0010042058989104892, "clip_ratio/low_min": 0.00027023089351132514, "clip_ratio/region_mean": 0.0023732945210213074, "completions/clipped_ratio": 0.0, "completions/max_length": 562.8, "completions/max_terminated_length": 562.8, "completions/mean_length": 203.1908920288086, "completions/mean_terminated_length": 203.1908920288086, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.02930880416339884, "epoch": 22.27119681665136, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3724061360.0, "reward": 8.304867506027222, "reward_std": 0.3492258310317993, "rewards/combined_reward/mean": 8.304867267608643, "rewards/combined_reward/std": 1.9514357805252076, "step": 36390 }, { "clip_ratio/high_max": 0.0025013101284760827, "clip_ratio/high_mean": 0.001368148250306452, "clip_ratio/low_mean": 0.0009073371892251695, "clip_ratio/low_min": 0.0002073685041978024, "clip_ratio/region_mean": 0.0022754854272837594, "completions/clipped_ratio": 0.0, "completions/max_length": 567.3, "completions/max_terminated_length": 567.3, "completions/mean_length": 202.06693115234376, "completions/mean_terminated_length": 202.06693115234376, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02802051668210576, "epoch": 22.28956228956229, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3727137457.0, "reward": 8.111579751968383, "reward_std": 0.37529766112565993, "rewards/combined_reward/mean": 8.11157946586609, "rewards/combined_reward/std": 2.1671497106552122, "step": 36420 }, { "clip_ratio/high_max": 0.002413817378449797, "clip_ratio/high_mean": 0.0013198624323801293, "clip_ratio/low_mean": 0.0008131112583214418, "clip_ratio/low_min": 0.00014613243717273388, "clip_ratio/region_mean": 0.0021329736984625923, "completions/clipped_ratio": 0.0, "completions/max_length": 592.4, "completions/max_terminated_length": 592.4, "completions/mean_length": 187.06302795410156, "completions/mean_terminated_length": 187.06302795410156, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.027004400896839798, "epoch": 22.307927762473216, "frac_reward_zero_std": 0.19583333916962148, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3730155939.0, "reward": 8.23948245048523, "reward_std": 0.3000768184661865, "rewards/combined_reward/mean": 8.239481973648072, "rewards/combined_reward/std": 2.003251779079437, "step": 36450 }, { "clip_ratio/high_max": 0.0026902361350948922, "clip_ratio/high_mean": 0.0014116223321858949, "clip_ratio/low_mean": 0.0009814178311595848, "clip_ratio/low_min": 0.0003091318467340898, "clip_ratio/region_mean": 0.00239304016965131, "completions/clipped_ratio": 0.0, "completions/max_length": 577.3, "completions/max_terminated_length": 577.3, "completions/mean_length": 203.69818420410155, "completions/mean_terminated_length": 203.69818420410155, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.03094184834820529, "epoch": 22.326293235384146, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3733238300.0, "reward": 7.895224285125733, "reward_std": 0.32241485416889193, "rewards/combined_reward/mean": 7.8952240467071535, "rewards/combined_reward/std": 2.54208961725235, "step": 36480 }, { "clip_ratio/high_max": 0.0026672055556749303, "clip_ratio/high_mean": 0.001482108045456698, "clip_ratio/low_mean": 0.0009170498970585565, "clip_ratio/low_min": 0.00022540830420136141, "clip_ratio/region_mean": 0.0023991579151091476, "completions/clipped_ratio": 0.0, "completions/max_length": 496.5, "completions/max_terminated_length": 496.5, "completions/mean_length": 199.87448425292968, "completions/mean_terminated_length": 199.87448425292968, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02873465685794751, "epoch": 22.34465870829507, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.039794921875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3736305978.0, "reward": 8.059447526931763, "reward_std": 0.29055422693490984, "rewards/combined_reward/mean": 8.059447431564331, "rewards/combined_reward/std": 2.4047374606132506, "step": 36510 }, { "clip_ratio/high_max": 0.002808304943027906, "clip_ratio/high_mean": 0.0014164247154743255, "clip_ratio/low_mean": 0.000945551323396406, "clip_ratio/low_min": 0.0003347454354904282, "clip_ratio/region_mean": 0.002361976008008545, "completions/clipped_ratio": 0.0, "completions/max_length": 624.7, "completions/max_terminated_length": 624.7, "completions/mean_length": 199.69818115234375, "completions/mean_terminated_length": 199.69818115234375, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.02896620302150647, "epoch": 22.363024181206, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3739372979.0, "reward": 8.174489068984986, "reward_std": 0.33731515109539034, "rewards/combined_reward/mean": 8.174489164352417, "rewards/combined_reward/std": 2.057419741153717, "step": 36540 }, { "clip_ratio/high_max": 0.002475455501795902, "clip_ratio/high_mean": 0.0015119745502791678, "clip_ratio/low_mean": 0.00097759730827723, "clip_ratio/low_min": 0.000311540223628981, "clip_ratio/region_mean": 0.002489571848612589, "completions/clipped_ratio": 0.0, "completions/max_length": 547.9, "completions/max_terminated_length": 547.9, "completions/mean_length": 204.70026702880858, "completions/mean_terminated_length": 204.70026702880858, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.030472164663175743, "epoch": 22.381389654116926, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3742459188.0, "reward": 8.254994106292724, "reward_std": 0.33534351363778114, "rewards/combined_reward/mean": 8.254993915557861, "rewards/combined_reward/std": 2.0233826637268066, "step": 36570 }, { "clip_ratio/high_max": 0.0025657850589292743, "clip_ratio/high_mean": 0.0015001453433796997, "clip_ratio/low_mean": 0.0010134914168399216, "clip_ratio/low_min": 0.0002821151756506879, "clip_ratio/region_mean": 0.0025136367626449403, "completions/clipped_ratio": 0.0, "completions/max_length": 649.1, "completions/max_terminated_length": 649.1, "completions/mean_length": 196.9565170288086, "completions/mean_terminated_length": 196.9565170288086, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.02786557961565753, "epoch": 22.399755127027856, "frac_reward_zero_std": 0.22500001043081283, "grad_norm": 0.0106201171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3745515661.0, "reward": 7.984451961517334, "reward_std": 0.28477169424295423, "rewards/combined_reward/mean": 7.984451723098755, "rewards/combined_reward/std": 2.2547364592552186, "step": 36600 }, { "clip_ratio/high_max": 0.002465224058929986, "clip_ratio/high_mean": 0.0014112845768977423, "clip_ratio/low_mean": 0.0009526130721496884, "clip_ratio/low_min": 0.000284473992845354, "clip_ratio/region_mean": 0.0023638976427416006, "completions/clipped_ratio": 0.0, "completions/max_length": 614.4, "completions/max_terminated_length": 614.4, "completions/mean_length": 202.9000045776367, "completions/mean_terminated_length": 202.9000045776367, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.030168424639850856, "epoch": 22.41812059993878, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3748594957.0, "reward": 7.987033557891846, "reward_std": 0.3381257578730583, "rewards/combined_reward/mean": 7.987033462524414, "rewards/combined_reward/std": 2.352073585987091, "step": 36630 }, { "clip_ratio/high_max": 0.0019805807334099274, "clip_ratio/high_mean": 0.0011505156677837174, "clip_ratio/low_mean": 0.0007042938284333407, "clip_ratio/low_min": 0.00018581513892665196, "clip_ratio/region_mean": 0.0018548095004613666, "completions/clipped_ratio": 0.0, "completions/max_length": 541.6, "completions/max_terminated_length": 541.6, "completions/mean_length": 192.25599212646483, "completions/mean_terminated_length": 192.25599212646483, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.025485574182433388, "epoch": 22.43648607284971, "frac_reward_zero_std": 0.16250000186264515, "grad_norm": 0.0096435546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3751633380.0, "reward": 8.244792413711547, "reward_std": 0.3392806351184845, "rewards/combined_reward/mean": 8.244792318344116, "rewards/combined_reward/std": 2.180273342132568, "step": 36660 }, { "clip_ratio/high_max": 0.002199698885669932, "clip_ratio/high_mean": 0.0012626496155765685, "clip_ratio/low_mean": 0.0008561350679277287, "clip_ratio/low_min": 0.0002132418626085079, "clip_ratio/region_mean": 0.0021187846835042973, "completions/clipped_ratio": 0.0, "completions/max_length": 560.2, "completions/max_terminated_length": 560.2, "completions/mean_length": 201.34089050292968, "completions/mean_terminated_length": 201.34089050292968, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.0266940992558375, "epoch": 22.454851545760636, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.028564453125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3754706689.0, "reward": 8.11965036392212, "reward_std": 0.36967952847480773, "rewards/combined_reward/mean": 8.119650173187257, "rewards/combined_reward/std": 2.3538220643997194, "step": 36690 }, { "clip_ratio/high_max": 0.002233866412764958, "clip_ratio/high_mean": 0.0012886406319012166, "clip_ratio/low_mean": 0.0008459970411422546, "clip_ratio/low_min": 0.0002095460785009588, "clip_ratio/region_mean": 0.002134637664615487, "completions/clipped_ratio": 0.0, "completions/max_length": 542.8, "completions/max_terminated_length": 542.8, "completions/mean_length": 197.33542175292968, "completions/mean_terminated_length": 197.33542175292968, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.02786512052019437, "epoch": 22.473217018671566, "frac_reward_zero_std": 0.19166667014360428, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3757764617.0, "reward": 7.964749574661255, "reward_std": 0.27330783009529114, "rewards/combined_reward/mean": 7.9647492408752445, "rewards/combined_reward/std": 2.3706244826316833, "step": 36720 }, { "clip_ratio/high_max": 0.0023402160450738544, "clip_ratio/high_mean": 0.0012806241369010725, "clip_ratio/low_mean": 0.0010304704080529822, "clip_ratio/low_min": 0.00031289908874896353, "clip_ratio/region_mean": 0.0023110945524725444, "completions/clipped_ratio": 0.0, "completions/max_length": 599.9, "completions/max_terminated_length": 599.9, "completions/mean_length": 200.79427490234374, "completions/mean_terminated_length": 200.79427490234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028023064392618834, "epoch": 22.49158249158249, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3760835827.0, "reward": 8.04759602546692, "reward_std": 0.3618509292602539, "rewards/combined_reward/mean": 8.047595882415772, "rewards/combined_reward/std": 2.2447476506233217, "step": 36750 }, { "clip_ratio/high_max": 0.0023286860712687487, "clip_ratio/high_mean": 0.0013305688788629291, "clip_ratio/low_mean": 0.0009309989770069175, "clip_ratio/low_min": 0.00023203221232203456, "clip_ratio/region_mean": 0.002261567826887282, "completions/clipped_ratio": 0.0, "completions/max_length": 497.1, "completions/max_terminated_length": 497.1, "completions/mean_length": 198.82500305175782, "completions/mean_terminated_length": 198.82500305175782, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.025453597058852515, "epoch": 22.50994796449342, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.025634765625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3763899475.0, "reward": 8.115602684020995, "reward_std": 0.3352783530950546, "rewards/combined_reward/mean": 8.115602445602416, "rewards/combined_reward/std": 2.4213062167167663, "step": 36780 }, { "clip_ratio/high_max": 0.0022934456164269553, "clip_ratio/high_mean": 0.0013950540009924832, "clip_ratio/low_mean": 0.0010398551678614846, "clip_ratio/low_min": 0.00028721762331163824, "clip_ratio/region_mean": 0.002434909189469181, "completions/clipped_ratio": 0.0, "completions/max_length": 569.7, "completions/max_terminated_length": 569.7, "completions/mean_length": 205.60833892822265, "completions/mean_terminated_length": 205.60833892822265, "completions/min_length": 45.3, "completions/min_terminated_length": 45.3, "entropy": 0.027859770444532236, "epoch": 22.528313437404346, "frac_reward_zero_std": 0.22083334140479566, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3766989171.0, "reward": 8.097163152694701, "reward_std": 0.27229110449552535, "rewards/combined_reward/mean": 8.097162914276122, "rewards/combined_reward/std": 2.078864312171936, "step": 36810 }, { "clip_ratio/high_max": 0.0026141866963977616, "clip_ratio/high_mean": 0.0014465364006658395, "clip_ratio/low_mean": 0.0009788120769978073, "clip_ratio/low_min": 0.0002708437025527625, "clip_ratio/region_mean": 0.0024253484662040137, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 200.54453735351564, "completions/mean_terminated_length": 200.54453735351564, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.03185436393444737, "epoch": 22.546678910315276, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3770059422.0, "reward": 8.070193815231324, "reward_std": 0.30895276367664337, "rewards/combined_reward/mean": 8.070193529129028, "rewards/combined_reward/std": 2.3244720458984376, "step": 36840 }, { "clip_ratio/high_max": 0.002566775687834403, "clip_ratio/high_mean": 0.0014668024336666956, "clip_ratio/low_mean": 0.001087250567919303, "clip_ratio/low_min": 0.00025745635187680215, "clip_ratio/region_mean": 0.002554052992976115, "completions/clipped_ratio": 0.0, "completions/max_length": 534.4, "completions/max_terminated_length": 534.4, "completions/mean_length": 193.5695373535156, "completions/mean_terminated_length": 193.5695373535156, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.030484908431147536, "epoch": 22.5650443832262, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3773102889.0, "reward": 8.052650690078735, "reward_std": 0.2877472579479218, "rewards/combined_reward/mean": 8.05265040397644, "rewards/combined_reward/std": 2.3068215370178224, "step": 36870 }, { "clip_ratio/high_max": 0.002291169483457149, "clip_ratio/high_mean": 0.0012682588344129424, "clip_ratio/low_mean": 0.0008813708440963334, "clip_ratio/low_min": 0.0002411835637758486, "clip_ratio/region_mean": 0.0021496296671102756, "completions/clipped_ratio": 0.0, "completions/max_length": 528.4, "completions/max_terminated_length": 528.4, "completions/mean_length": 183.74349517822264, "completions/mean_terminated_length": 183.74349517822264, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02701814384199679, "epoch": 22.58340985613713, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 3776108624.0, "reward": 7.896353149414063, "reward_std": 0.2618178591132164, "rewards/combined_reward/mean": 7.896352767944336, "rewards/combined_reward/std": 2.5572125792503355, "step": 36900 }, { "clip_ratio/high_max": 0.0028976731865744416, "clip_ratio/high_mean": 0.0015614871134554656, "clip_ratio/low_mean": 0.0010179229094016288, "clip_ratio/low_min": 0.000271549206081545, "clip_ratio/region_mean": 0.002579409993389466, "completions/clipped_ratio": 0.0, "completions/max_length": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 205.5679733276367, "completions/mean_terminated_length": 205.5679733276367, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.03163683004677296, "epoch": 22.601775329048056, "frac_reward_zero_std": 0.14166666902601718, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3779198165.0, "reward": 7.986977624893188, "reward_std": 0.33562025129795076, "rewards/combined_reward/mean": 7.9869773387908936, "rewards/combined_reward/std": 2.3393388867378233, "step": 36930 }, { "clip_ratio/high_max": 0.002215165823145071, "clip_ratio/high_mean": 0.0012571465917668927, "clip_ratio/low_mean": 0.001236642371683653, "clip_ratio/low_min": 0.000276881529862294, "clip_ratio/region_mean": 0.002493788977638663, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 186.4947967529297, "completions/mean_terminated_length": 186.4947967529297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.031114380434155466, "epoch": 22.620140801958982, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 3782214465.0, "reward": 8.044262552261353, "reward_std": 0.38777360022068025, "rewards/combined_reward/mean": 8.044262361526489, "rewards/combined_reward/std": 2.0346168994903566, "step": 36960 }, { "clip_ratio/high_max": 0.0025405740375087286, "clip_ratio/high_mean": 0.001430290864785396, "clip_ratio/low_mean": 0.00110451089591758, "clip_ratio/low_min": 0.0004221314018650446, "clip_ratio/region_mean": 0.002534801738996369, "completions/clipped_ratio": 0.0, "completions/max_length": 556.7, "completions/max_terminated_length": 556.7, "completions/mean_length": 202.4252670288086, "completions/mean_terminated_length": 202.4252670288086, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.0315357037819922, "epoch": 22.63850627486991, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.00775146484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3785291938.0, "reward": 8.139716625213623, "reward_std": 0.3273388832807541, "rewards/combined_reward/mean": 8.139716243743896, "rewards/combined_reward/std": 2.2314311623573304, "step": 36990 }, { "clip_ratio/high_max": 0.0022896791939274407, "clip_ratio/high_mean": 0.001361382217388988, "clip_ratio/low_mean": 0.0010614099628583062, "clip_ratio/low_min": 0.0003226220653838633, "clip_ratio/region_mean": 0.002422792166180443, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 200.7612045288086, "completions/mean_terminated_length": 200.7612045288086, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.03238465079727272, "epoch": 22.656871747780837, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3788363021.0, "reward": 7.881092882156372, "reward_std": 0.33853748738765715, "rewards/combined_reward/mean": 7.881092500686646, "rewards/combined_reward/std": 2.3908644318580627, "step": 37020 }, { "clip_ratio/high_max": 0.0026961284151184373, "clip_ratio/high_mean": 0.0014353591747446142, "clip_ratio/low_mean": 0.0010102980758044092, "clip_ratio/low_min": 0.00033305417528026735, "clip_ratio/region_mean": 0.002445657240362683, "completions/clipped_ratio": 0.0, "completions/max_length": 581.7, "completions/max_terminated_length": 581.7, "completions/mean_length": 198.49844207763672, "completions/mean_terminated_length": 198.49844207763672, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03280628143499295, "epoch": 22.675237220691766, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3791425415.0, "reward": 8.132589435577392, "reward_std": 0.3336997151374817, "rewards/combined_reward/mean": 8.132589101791382, "rewards/combined_reward/std": 2.063267207145691, "step": 37050 }, { "clip_ratio/high_max": 0.002683416386328948, "clip_ratio/high_mean": 0.0014719877023405086, "clip_ratio/low_mean": 0.0009957561145711225, "clip_ratio/low_min": 0.0002734203543999077, "clip_ratio/region_mean": 0.0024677437999343966, "completions/clipped_ratio": 0.0, "completions/max_length": 548.9, "completions/max_terminated_length": 548.9, "completions/mean_length": 187.56146240234375, "completions/mean_terminated_length": 187.56146240234375, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030915874956796565, "epoch": 22.693602693602692, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 3794445811.0, "reward": 8.024460697174073, "reward_std": 0.32348952889442445, "rewards/combined_reward/mean": 8.024460601806641, "rewards/combined_reward/std": 2.372605526447296, "step": 37080 }, { "clip_ratio/high_max": 0.002378848840695961, "clip_ratio/high_mean": 0.0013156185080636835, "clip_ratio/low_mean": 0.0008913054651202401, "clip_ratio/low_min": 0.00029994198193890044, "clip_ratio/region_mean": 0.002206923984825456, "completions/clipped_ratio": 0.0, "completions/max_length": 541.7, "completions/max_terminated_length": 541.7, "completions/mean_length": 202.63828887939454, "completions/mean_terminated_length": 202.63828887939454, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029680356841223936, "epoch": 22.71196816651362, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3797524102.0, "reward": 7.893038749694824, "reward_std": 0.2942368075251579, "rewards/combined_reward/mean": 7.893038511276245, "rewards/combined_reward/std": 2.4786084651947022, "step": 37110 }, { "clip_ratio/high_max": 0.0024433668765899102, "clip_ratio/high_mean": 0.0013492324703596144, "clip_ratio/low_mean": 0.0008743120215513045, "clip_ratio/low_min": 0.0002260652773353892, "clip_ratio/region_mean": 0.0022235445081605575, "completions/clipped_ratio": 0.0, "completions/max_length": 571.7, "completions/max_terminated_length": 571.7, "completions/mean_length": 200.29141387939453, "completions/mean_terminated_length": 200.29141387939453, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.029988353978842495, "epoch": 22.730333639424547, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3800593381.0, "reward": 8.1971999168396, "reward_std": 0.3091407150030136, "rewards/combined_reward/mean": 8.197199630737305, "rewards/combined_reward/std": 1.9524265646934509, "step": 37140 }, { "clip_ratio/high_max": 0.003052991828978217, "clip_ratio/high_mean": 0.0016701702552381903, "clip_ratio/low_mean": 0.0009494949291062464, "clip_ratio/low_min": 0.00030862572020851076, "clip_ratio/region_mean": 0.002619665193196852, "completions/clipped_ratio": 0.0, "completions/max_length": 618.9, "completions/max_terminated_length": 618.9, "completions/mean_length": 205.2135452270508, "completions/mean_terminated_length": 205.2135452270508, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03286903040328373, "epoch": 22.748699112335476, "frac_reward_zero_std": 0.1833333358168602, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3803681561.0, "reward": 7.565158462524414, "reward_std": 0.3231548830866814, "rewards/combined_reward/mean": 7.565158367156982, "rewards/combined_reward/std": 2.8443074226379395, "step": 37170 }, { "clip_ratio/high_max": 0.0023378862754422397, "clip_ratio/high_mean": 0.001326057060820555, "clip_ratio/low_mean": 0.0009884589107969077, "clip_ratio/low_min": 0.00036909275562114395, "clip_ratio/region_mean": 0.002314515992475208, "completions/clipped_ratio": 0.0, "completions/max_length": 682.2, "completions/max_terminated_length": 682.2, "completions/mean_length": 191.4143280029297, "completions/mean_terminated_length": 191.4143280029297, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02953113168478012, "epoch": 22.767064585246402, "frac_reward_zero_std": 0.21250000447034836, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3806716752.0, "reward": 8.128869152069091, "reward_std": 0.3075192704796791, "rewards/combined_reward/mean": 8.128868770599365, "rewards/combined_reward/std": 2.285439157485962, "step": 37200 }, { "clip_ratio/high_max": 0.0025489584006815375, "clip_ratio/high_mean": 0.0014547090904670767, "clip_ratio/low_mean": 0.001025189465387181, "clip_ratio/low_min": 0.00033596559490736884, "clip_ratio/region_mean": 0.0024798985565818534, "completions/clipped_ratio": 0.0, "completions/max_length": 559.8, "completions/max_terminated_length": 559.8, "completions/mean_length": 202.71198425292968, "completions/mean_terminated_length": 202.71198425292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.0291947563799719, "epoch": 22.78543005815733, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3809795326.0, "reward": 8.12508864402771, "reward_std": 0.3139971770346165, "rewards/combined_reward/mean": 8.125088596343994, "rewards/combined_reward/std": 1.9977913200855255, "step": 37230 }, { "clip_ratio/high_max": 0.002370326285502718, "clip_ratio/high_mean": 0.001431231978252375, "clip_ratio/low_mean": 0.0008176069926169779, "clip_ratio/low_min": 0.00021645180798562553, "clip_ratio/region_mean": 0.0022488389508604693, "completions/clipped_ratio": 0.0, "completions/max_length": 578.2, "completions/max_terminated_length": 578.2, "completions/mean_length": 199.4627655029297, "completions/mean_terminated_length": 199.4627655029297, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03043824991521736, "epoch": 22.803795531068257, "frac_reward_zero_std": 0.1500000052154064, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3812861423.0, "reward": 7.961273050308227, "reward_std": 0.39496155381202697, "rewards/combined_reward/mean": 7.961272716522217, "rewards/combined_reward/std": 2.206748938560486, "step": 37260 }, { "clip_ratio/high_max": 0.0029339465817126136, "clip_ratio/high_mean": 0.0016212840642159184, "clip_ratio/low_mean": 0.0009592444331550117, "clip_ratio/low_min": 0.00022651901235803962, "clip_ratio/region_mean": 0.0025805285025853665, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 209.72604522705078, "completions/mean_terminated_length": 209.72604522705078, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03148927010285358, "epoch": 22.822161003979186, "frac_reward_zero_std": 0.15000000223517418, "grad_norm": 0.0263671875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3815966931.0, "reward": 7.898429441452026, "reward_std": 0.4162068605422974, "rewards/combined_reward/mean": 7.898429298400879, "rewards/combined_reward/std": 2.4188475966453553, "step": 37290 }, { "clip_ratio/high_max": 0.001923817638453329, "clip_ratio/high_mean": 0.0010965601450152462, "clip_ratio/low_mean": 0.0008452380752108486, "clip_ratio/low_min": 0.0002716788454563357, "clip_ratio/region_mean": 0.0019417982364151006, "completions/clipped_ratio": 0.0, "completions/max_length": 586.7, "completions/max_terminated_length": 586.7, "completions/mean_length": 192.6109420776367, "completions/mean_terminated_length": 192.6109420776367, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.02396102943457663, "epoch": 22.840526476890112, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0084228515625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3819006717.0, "reward": 7.914506435394287, "reward_std": 0.29446312189102175, "rewards/combined_reward/mean": 7.914506244659424, "rewards/combined_reward/std": 2.3049142479896547, "step": 37320 }, { "clip_ratio/high_max": 0.0022165131641183203, "clip_ratio/high_mean": 0.00129077244613048, "clip_ratio/low_mean": 0.0008815383338514948, "clip_ratio/low_min": 0.0002460415632716225, "clip_ratio/region_mean": 0.0021723107958678154, "completions/clipped_ratio": 0.0, "completions/max_length": 640.6, "completions/max_terminated_length": 640.6, "completions/mean_length": 208.91380767822267, "completions/mean_terminated_length": 208.91380767822267, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.02660215344900886, "epoch": 22.85889194980104, "frac_reward_zero_std": 0.14166666902601718, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3822109106.0, "reward": 7.909080600738525, "reward_std": 0.34108992740511895, "rewards/combined_reward/mean": 7.90908031463623, "rewards/combined_reward/std": 2.3730401515960695, "step": 37350 }, { "clip_ratio/high_max": 0.0024609605374280364, "clip_ratio/high_mean": 0.0013587834712476857, "clip_ratio/low_mean": 0.0009597419822966913, "clip_ratio/low_min": 0.0002156446588439091, "clip_ratio/region_mean": 0.002318525445298292, "completions/clipped_ratio": 0.0, "completions/max_length": 621.5, "completions/max_terminated_length": 621.5, "completions/mean_length": 198.74036865234376, "completions/mean_terminated_length": 198.74036865234376, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030199648082877197, "epoch": 22.877257422711967, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3825172429.0, "reward": 8.038813924789428, "reward_std": 0.371012008190155, "rewards/combined_reward/mean": 8.038813591003418, "rewards/combined_reward/std": 2.3880921840667724, "step": 37380 }, { "clip_ratio/high_max": 0.002183766835757221, "clip_ratio/high_mean": 0.001307453239132883, "clip_ratio/low_mean": 0.0009305328715223974, "clip_ratio/low_min": 0.00024057055998127908, "clip_ratio/region_mean": 0.002237986123509472, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 212.44037017822265, "completions/mean_terminated_length": 212.44037017822265, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.029204878862947224, "epoch": 22.895622895622896, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0076904296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3828288360.0, "reward": 8.027178192138672, "reward_std": 0.36859096884727477, "rewards/combined_reward/mean": 8.027178049087524, "rewards/combined_reward/std": 2.2747610449790954, "step": 37410 }, { "clip_ratio/high_max": 0.002210841644409811, "clip_ratio/high_mean": 0.0013546492317497419, "clip_ratio/low_mean": 0.0009898321734605512, "clip_ratio/low_min": 0.00029792771165375597, "clip_ratio/region_mean": 0.002344481421459932, "completions/clipped_ratio": 0.0, "completions/max_length": 572.8, "completions/max_terminated_length": 572.8, "completions/mean_length": 210.57890930175782, "completions/mean_terminated_length": 210.57890930175782, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.028587921289727092, "epoch": 22.913988368533822, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0057373046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3831397143.0, "reward": 7.954912185668945, "reward_std": 0.3482479453086853, "rewards/combined_reward/mean": 7.954911947250366, "rewards/combined_reward/std": 2.296778988838196, "step": 37440 }, { "clip_ratio/high_max": 0.0022970813389595907, "clip_ratio/high_mean": 0.0013107176239524657, "clip_ratio/low_mean": 0.000949608582353297, "clip_ratio/low_min": 0.00022863520037693283, "clip_ratio/region_mean": 0.0022603262278911036, "completions/clipped_ratio": 0.0, "completions/max_length": 590.6, "completions/max_terminated_length": 590.6, "completions/mean_length": 202.18594055175782, "completions/mean_terminated_length": 202.18594055175782, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.027306721716498335, "epoch": 22.93235384144475, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3834473697.0, "reward": 8.187969732284547, "reward_std": 0.32823544144630434, "rewards/combined_reward/mean": 8.187969398498534, "rewards/combined_reward/std": 2.1979069113731384, "step": 37470 }, { "clip_ratio/high_max": 0.002566274371444403, "clip_ratio/high_mean": 0.0015008446634359037, "clip_ratio/low_mean": 0.0011312572631140938, "clip_ratio/low_min": 0.0004117700397425021, "clip_ratio/region_mean": 0.0026321019152722633, "completions/clipped_ratio": 0.0, "completions/max_length": 587.6, "completions/max_terminated_length": 587.6, "completions/mean_length": 210.30912017822266, "completions/mean_terminated_length": 210.30912017822266, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.032141021359711884, "epoch": 22.950719314355677, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0255126953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3837581444.0, "reward": 7.837403964996338, "reward_std": 0.2984386160969734, "rewards/combined_reward/mean": 7.837403869628906, "rewards/combined_reward/std": 2.4556063175201417, "step": 37500 }, { "clip_ratio/high_max": 0.002462958442629315, "clip_ratio/high_mean": 0.0014858188029999535, "clip_ratio/low_mean": 0.0009337762364642307, "clip_ratio/low_min": 0.0002881682038908669, "clip_ratio/region_mean": 0.002419595045163684, "completions/clipped_ratio": 0.0, "completions/max_length": 542.5, "completions/max_terminated_length": 542.5, "completions/mean_length": 203.5666717529297, "completions/mean_terminated_length": 203.5666717529297, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.03130309864257773, "epoch": 22.969084787266606, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 3840663300.0, "reward": 7.971546363830567, "reward_std": 0.3404989555478096, "rewards/combined_reward/mean": 7.971545886993408, "rewards/combined_reward/std": 2.3441709876060486, "step": 37530 }, { "clip_ratio/high_max": 0.0023318181636568626, "clip_ratio/high_mean": 0.0012644593800359872, "clip_ratio/low_mean": 0.0009478162320495661, "clip_ratio/low_min": 0.0002560387261231275, "clip_ratio/region_mean": 0.0022122756044457978, "completions/clipped_ratio": 0.0, "completions/max_length": 540.8, "completions/max_terminated_length": 540.8, "completions/mean_length": 190.9127670288086, "completions/mean_terminated_length": 190.9127670288086, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.02851511163947483, "epoch": 22.987450260177532, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.04248046875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3843696565.0, "reward": 8.104946899414063, "reward_std": 0.35407122522592543, "rewards/combined_reward/mean": 8.104946517944336, "rewards/combined_reward/std": 2.423997676372528, "step": 37560 }, { "clip_ratio/high_max": 0.0024021703254636694, "clip_ratio/high_mean": 0.0014825245536670437, "clip_ratio/low_mean": 0.0008681226194682295, "clip_ratio/low_min": 0.00022930010139026632, "clip_ratio/region_mean": 0.00235064720470553, "completions/clipped_ratio": 0.0, "completions/max_length": 536.7, "completions/max_terminated_length": 536.7, "completions/mean_length": 198.15912170410155, "completions/mean_terminated_length": 198.15912170410155, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.03232745534219479, "epoch": 23.005509641873278, "frac_reward_zero_std": 0.15833333767950536, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 3846757656.0, "reward": 8.248012065887451, "reward_std": 0.3513083398342133, "rewards/combined_reward/mean": 8.248011779785156, "rewards/combined_reward/std": 2.0222553491592405, "step": 37590 }, { "clip_ratio/high_max": 0.0021410385287405614, "clip_ratio/high_mean": 0.0011590879268624122, "clip_ratio/low_mean": 0.0008098406635933012, "clip_ratio/low_min": 0.00029197708524103896, "clip_ratio/region_mean": 0.0019689286059777563, "completions/clipped_ratio": 0.0, "completions/max_length": 597.1, "completions/max_terminated_length": 597.1, "completions/mean_length": 195.25781707763673, "completions/mean_terminated_length": 195.25781707763673, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03067738449511429, "epoch": 23.023875114784207, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3849807606.0, "reward": 7.954011487960815, "reward_std": 0.31704982817173005, "rewards/combined_reward/mean": 7.9540114402771, "rewards/combined_reward/std": 2.3888933181762697, "step": 37620 }, { "clip_ratio/high_max": 0.0019805666248430497, "clip_ratio/high_mean": 0.0010898059822163002, "clip_ratio/low_mean": 0.0007573261222811804, "clip_ratio/low_min": 0.00021980840053098898, "clip_ratio/region_mean": 0.0018471321047400125, "completions/clipped_ratio": 0.0, "completions/max_length": 597.4, "completions/max_terminated_length": 597.4, "completions/mean_length": 204.5291717529297, "completions/mean_terminated_length": 204.5291717529297, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.029595667496323586, "epoch": 23.042240587695133, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.0057373046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3852893158.0, "reward": 7.880979061126709, "reward_std": 0.3626251742243767, "rewards/combined_reward/mean": 7.880978441238403, "rewards/combined_reward/std": 2.2495171785354615, "step": 37650 }, { "clip_ratio/high_max": 0.002079101770262544, "clip_ratio/high_mean": 0.0011377837351271107, "clip_ratio/low_mean": 0.0007764673951896839, "clip_ratio/low_min": 0.0002640970832241389, "clip_ratio/region_mean": 0.0019142511426859224, "completions/clipped_ratio": 0.0, "completions/max_length": 547.3, "completions/max_terminated_length": 547.3, "completions/mean_length": 190.76615142822266, "completions/mean_terminated_length": 190.76615142822266, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.031135561891521016, "epoch": 23.060606060606062, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3855925860.0, "reward": 7.908791971206665, "reward_std": 0.35010786652565, "rewards/combined_reward/mean": 7.908791780471802, "rewards/combined_reward/std": 2.1953639507293703, "step": 37680 }, { "clip_ratio/high_max": 0.0020310982280837683, "clip_ratio/high_mean": 0.0011613436513774407, "clip_ratio/low_mean": 0.0009335817010044896, "clip_ratio/low_min": 0.0003106489702380107, "clip_ratio/region_mean": 0.0020949253691166327, "completions/clipped_ratio": 0.0, "completions/max_length": 542.7, "completions/max_terminated_length": 542.7, "completions/mean_length": 206.74505767822265, "completions/mean_terminated_length": 206.74505767822265, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.03070769631303847, "epoch": 23.078971533516988, "frac_reward_zero_std": 0.17083333991467953, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3859019921.0, "reward": 8.263886499404908, "reward_std": 0.3712050586938858, "rewards/combined_reward/mean": 8.263886547088623, "rewards/combined_reward/std": 2.2222760677337647, "step": 37710 }, { "clip_ratio/high_max": 0.001796658301949113, "clip_ratio/high_mean": 0.001065659107613707, "clip_ratio/low_mean": 0.0007262602595195252, "clip_ratio/low_min": 0.00021983634360367432, "clip_ratio/region_mean": 0.001791919371074376, "completions/clipped_ratio": 0.0, "completions/max_length": 603.4, "completions/max_terminated_length": 603.4, "completions/mean_length": 202.08099670410155, "completions/mean_terminated_length": 202.08099670410155, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028342823218554258, "epoch": 23.097337006427914, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.00494384765625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3862096072.0, "reward": 8.037287664413451, "reward_std": 0.3167830675840378, "rewards/combined_reward/mean": 8.037287473678589, "rewards/combined_reward/std": 2.3369547724723816, "step": 37740 }, { "clip_ratio/high_max": 0.0019271978012208517, "clip_ratio/high_mean": 0.0011526080679080527, "clip_ratio/low_mean": 0.0008856332717793217, "clip_ratio/low_min": 0.00024593074752677545, "clip_ratio/region_mean": 0.002038241359211194, "completions/clipped_ratio": 0.0, "completions/max_length": 643.9, "completions/max_terminated_length": 643.9, "completions/mean_length": 203.62709045410156, "completions/mean_terminated_length": 203.62709045410156, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.031294043036177756, "epoch": 23.115702479338843, "frac_reward_zero_std": 0.15833333544433117, "grad_norm": 0.031494140625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3865178160.0, "reward": 7.980869626998901, "reward_std": 0.3927024468779564, "rewards/combined_reward/mean": 7.980869293212891, "rewards/combined_reward/std": 2.5331424117088317, "step": 37770 }, { "clip_ratio/high_max": 0.0018626579995422313, "clip_ratio/high_mean": 0.0011008521132074142, "clip_ratio/low_mean": 0.0007158845372032374, "clip_ratio/low_min": 0.0001735644965568402, "clip_ratio/region_mean": 0.0018167366370713959, "completions/clipped_ratio": 0.0, "completions/max_length": 566.8, "completions/max_terminated_length": 566.8, "completions/mean_length": 198.96953582763672, "completions/mean_terminated_length": 198.96953582763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027546531660482286, "epoch": 23.13406795224977, "frac_reward_zero_std": 0.17083333618938923, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3868242363.0, "reward": 7.927867841720581, "reward_std": 0.36569158881902697, "rewards/combined_reward/mean": 7.927867603302002, "rewards/combined_reward/std": 2.471423923969269, "step": 37800 }, { "clip_ratio/high_max": 0.0016763700788336186, "clip_ratio/high_mean": 0.0009480125208938261, "clip_ratio/low_mean": 0.0006637338945438386, "clip_ratio/low_min": 0.00021049982193896237, "clip_ratio/region_mean": 0.0016117463974903027, "completions/clipped_ratio": 0.0, "completions/max_length": 565.4, "completions/max_terminated_length": 565.4, "completions/mean_length": 200.08698425292968, "completions/mean_terminated_length": 200.08698425292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02632363000884652, "epoch": 23.152433425160698, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3871310857.0, "reward": 8.06325044631958, "reward_std": 0.32541149109601974, "rewards/combined_reward/mean": 8.063250255584716, "rewards/combined_reward/std": 2.2190611362457275, "step": 37830 }, { "clip_ratio/high_max": 0.0021483438055535467, "clip_ratio/high_mean": 0.0011740064060480411, "clip_ratio/low_mean": 0.0007546516631312746, "clip_ratio/low_min": 0.00019487863100948745, "clip_ratio/region_mean": 0.0019286580796081883, "completions/clipped_ratio": 0.0, "completions/max_length": 580.8, "completions/max_terminated_length": 580.8, "completions/mean_length": 199.7401107788086, "completions/mean_terminated_length": 199.7401107788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.030706776973480978, "epoch": 23.170798898071624, "frac_reward_zero_std": 0.2125000085681677, "grad_norm": 0.0240478515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3874378019.0, "reward": 7.886414480209351, "reward_std": 0.37768165916204455, "rewards/combined_reward/mean": 7.886414194107056, "rewards/combined_reward/std": 2.5516385316848753, "step": 37860 }, { "clip_ratio/high_max": 0.0020153147842696248, "clip_ratio/high_mean": 0.0011477498650492635, "clip_ratio/low_mean": 0.000781767197016355, "clip_ratio/low_min": 0.00020974632667882057, "clip_ratio/region_mean": 0.0019295170563661183, "completions/clipped_ratio": 0.0, "completions/max_length": 515.2, "completions/max_terminated_length": 515.2, "completions/mean_length": 199.81771392822264, "completions/mean_terminated_length": 199.81771392822264, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.028413771154979863, "epoch": 23.189164370982553, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.0076904296875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3877445479.0, "reward": 8.177104425430297, "reward_std": 0.2655133530497551, "rewards/combined_reward/mean": 8.177104043960572, "rewards/combined_reward/std": 1.914379107952118, "step": 37890 }, { "clip_ratio/high_max": 0.0019404042580087359, "clip_ratio/high_mean": 0.0011130002625577618, "clip_ratio/low_mean": 0.0007480095586288371, "clip_ratio/low_min": 0.00023367168153830183, "clip_ratio/region_mean": 0.0018610098093631677, "completions/clipped_ratio": 0.0, "completions/max_length": 598.7, "completions/max_terminated_length": 598.7, "completions/mean_length": 197.03724517822266, "completions/mean_terminated_length": 197.03724517822266, "completions/min_length": 48.7, "completions/min_terminated_length": 48.7, "entropy": 0.03133835276433577, "epoch": 23.20752984389348, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.007781982421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3880502262.0, "reward": 8.104157638549804, "reward_std": 0.3298863172531128, "rewards/combined_reward/mean": 8.10415735244751, "rewards/combined_reward/std": 2.2037118554115294, "step": 37920 }, { "clip_ratio/high_max": 0.0019040850442252123, "clip_ratio/high_mean": 0.001136597649504741, "clip_ratio/low_mean": 0.0008180726523278281, "clip_ratio/low_min": 0.0002673194176168181, "clip_ratio/region_mean": 0.0019546702998923136, "completions/clipped_ratio": 0.0, "completions/max_length": 506.3, "completions/max_terminated_length": 506.3, "completions/mean_length": 198.5533905029297, "completions/mean_terminated_length": 198.5533905029297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02942678698649009, "epoch": 23.225895316804408, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3883564867.0, "reward": 8.3661958694458, "reward_std": 0.32404734641313554, "rewards/combined_reward/mean": 8.366195487976075, "rewards/combined_reward/std": 1.8183851957321167, "step": 37950 }, { "clip_ratio/high_max": 0.001840141285113835, "clip_ratio/high_mean": 0.0011385144379649622, "clip_ratio/low_mean": 0.0008976037276928158, "clip_ratio/low_min": 0.00027105819074980296, "clip_ratio/region_mean": 0.0020361181760866504, "completions/clipped_ratio": 0.0, "completions/max_length": 621.7, "completions/max_terminated_length": 621.7, "completions/mean_length": 195.44479675292968, "completions/mean_terminated_length": 195.44479675292968, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.0308924412742878, "epoch": 23.244260789715334, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.004974365234375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 3886615535.0, "reward": 7.720830488204956, "reward_std": 0.33034035563468933, "rewards/combined_reward/mean": 7.720830249786377, "rewards/combined_reward/std": 2.4837602615356444, "step": 37980 }, { "clip_ratio/high_max": 0.001689579346566461, "clip_ratio/high_mean": 0.0009316517767729238, "clip_ratio/low_mean": 0.0007133062739740126, "clip_ratio/low_min": 0.00018398603424429893, "clip_ratio/region_mean": 0.0016449580463813618, "completions/clipped_ratio": 0.0, "completions/max_length": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 186.99414443969727, "completions/mean_terminated_length": 186.99414443969727, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.026685053878463805, "epoch": 23.262014080195897, "frac_reward_zero_std": 0.2812500074505806, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 3889710342.0, "reward": 7.9829113483428955, "reward_std": 0.2336299866437912, "rewards/combined_reward/mean": 7.982910871505737, "rewards/combined_reward/std": 2.308204472064972, "step": 38010 }, { "clip_ratio/high_max": 0.0020869382574649838, "clip_ratio/high_mean": 0.00120067137395381, "clip_ratio/low_mean": 0.000811462061938073, "clip_ratio/low_min": 0.0002484415747555128, "clip_ratio/region_mean": 0.002012133436316314, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 204.48646087646483, "completions/mean_terminated_length": 204.48646087646483, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030074963392689825, "epoch": 23.280379553106826, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3892795730.0, "reward": 8.037982654571532, "reward_std": 0.3246257096529007, "rewards/combined_reward/mean": 8.037982606887818, "rewards/combined_reward/std": 2.222760224342346, "step": 38040 }, { "clip_ratio/high_max": 0.0021353081096701014, "clip_ratio/high_mean": 0.0012144040408505437, "clip_ratio/low_mean": 0.000755990644756821, "clip_ratio/low_min": 0.00022161471982447742, "clip_ratio/region_mean": 0.001970394685728631, "completions/clipped_ratio": 0.0, "completions/max_length": 657.5, "completions/max_terminated_length": 657.5, "completions/mean_length": 210.43255615234375, "completions/mean_terminated_length": 210.43255615234375, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03197485947360595, "epoch": 23.29874502601775, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.0252685546875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 3895903951.0, "reward": 8.005599212646484, "reward_std": 0.3861128121614456, "rewards/combined_reward/mean": 8.00559892654419, "rewards/combined_reward/std": 2.2220882415771483, "step": 38070 }, { "clip_ratio/high_max": 0.0019625122852933905, "clip_ratio/high_mean": 0.0011845972891023847, "clip_ratio/low_mean": 0.0008190490695900129, "clip_ratio/low_min": 0.00025834194384515284, "clip_ratio/region_mean": 0.00200364635772227, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 194.22891235351562, "completions/mean_terminated_length": 194.22891235351562, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.028747437009587883, "epoch": 23.31711049892868, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0101318359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3898949950.0, "reward": 8.156330823898315, "reward_std": 0.3069500356912613, "rewards/combined_reward/mean": 8.156330728530884, "rewards/combined_reward/std": 2.1450107216835024, "step": 38100 }, { "clip_ratio/high_max": 0.0020431121098226867, "clip_ratio/high_mean": 0.0011695016561134252, "clip_ratio/low_mean": 0.0007265876140688003, "clip_ratio/low_min": 0.0002425349916544898, "clip_ratio/region_mean": 0.0018960892615723425, "completions/clipped_ratio": 0.0, "completions/max_length": 527.6, "completions/max_terminated_length": 527.6, "completions/mean_length": 199.43594207763672, "completions/mean_terminated_length": 199.43594207763672, "completions/min_length": 51.8, "completions/min_terminated_length": 51.8, "entropy": 0.029626724123954774, "epoch": 23.335475971839607, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.029296875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3902015944.0, "reward": 8.108083057403565, "reward_std": 0.3561324506998062, "rewards/combined_reward/mean": 8.108082723617553, "rewards/combined_reward/std": 2.3205361366271973, "step": 38130 }, { "clip_ratio/high_max": 0.0019231723180079521, "clip_ratio/high_mean": 0.001020094592119373, "clip_ratio/low_mean": 0.0007563942980292874, "clip_ratio/low_min": 0.00019624198854823288, "clip_ratio/region_mean": 0.001776488888087139, "completions/clipped_ratio": 0.0, "completions/max_length": 681.8, "completions/max_terminated_length": 681.8, "completions/mean_length": 201.27474517822264, "completions/mean_terminated_length": 201.27474517822264, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029675133309016624, "epoch": 23.353841444750536, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3905088999.0, "reward": 8.01432671546936, "reward_std": 0.26132280826568605, "rewards/combined_reward/mean": 8.014327049255371, "rewards/combined_reward/std": 2.2462239265441895, "step": 38160 }, { "clip_ratio/high_max": 0.002336121692011754, "clip_ratio/high_mean": 0.0012651510420255362, "clip_ratio/low_mean": 0.0008420012936160977, "clip_ratio/low_min": 0.00028260482989329226, "clip_ratio/region_mean": 0.002107152333580113, "completions/clipped_ratio": 0.0, "completions/max_length": 505.7, "completions/max_terminated_length": 505.7, "completions/mean_length": 201.52787322998046, "completions/mean_terminated_length": 201.52787322998046, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03144517828089496, "epoch": 23.37220691766146, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3908163026.0, "reward": 8.20033564567566, "reward_std": 0.3363314375281334, "rewards/combined_reward/mean": 8.20033540725708, "rewards/combined_reward/std": 1.981030559539795, "step": 38190 }, { "clip_ratio/high_max": 0.001845576050497281, "clip_ratio/high_mean": 0.0010274827948402768, "clip_ratio/low_mean": 0.0008239906193921343, "clip_ratio/low_min": 0.000213592800234134, "clip_ratio/region_mean": 0.0018514734270866028, "completions/clipped_ratio": 0.0, "completions/max_length": 591.4, "completions/max_terminated_length": 591.4, "completions/mean_length": 202.25313262939454, "completions/mean_terminated_length": 202.25313262939454, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.026686821738258003, "epoch": 23.39057239057239, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3911239838.0, "reward": 8.11979513168335, "reward_std": 0.30522665828466417, "rewards/combined_reward/mean": 8.119794750213623, "rewards/combined_reward/std": 2.336755120754242, "step": 38220 }, { "clip_ratio/high_max": 0.002227132126184491, "clip_ratio/high_mean": 0.001232959088035083, "clip_ratio/low_mean": 0.0007362979026462805, "clip_ratio/low_min": 0.00019206208210865346, "clip_ratio/region_mean": 0.001969257002444162, "completions/clipped_ratio": 0.0, "completions/max_length": 628.5, "completions/max_terminated_length": 628.5, "completions/mean_length": 203.7765686035156, "completions/mean_terminated_length": 203.7765686035156, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03090629200451076, "epoch": 23.408937863483317, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3914322500.0, "reward": 8.291445684432983, "reward_std": 0.31071844324469566, "rewards/combined_reward/mean": 8.29144549369812, "rewards/combined_reward/std": 2.083682429790497, "step": 38250 }, { "clip_ratio/high_max": 0.0019319512978351364, "clip_ratio/high_mean": 0.001062784968841394, "clip_ratio/low_mean": 0.0008216117939203589, "clip_ratio/low_min": 0.00023433894724197066, "clip_ratio/region_mean": 0.001884396756455923, "completions/clipped_ratio": 0.0, "completions/max_length": 608.7, "completions/max_terminated_length": 608.7, "completions/mean_length": 205.7385482788086, "completions/mean_terminated_length": 205.7385482788086, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029121986706741154, "epoch": 23.427303336394246, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3917412696.0, "reward": 8.049729347229004, "reward_std": 0.3036227323114872, "rewards/combined_reward/mean": 8.049728870391846, "rewards/combined_reward/std": 2.2800207138061523, "step": 38280 }, { "clip_ratio/high_max": 0.0022617913028322315, "clip_ratio/high_mean": 0.0013007670944110334, "clip_ratio/low_mean": 0.0008247537166729065, "clip_ratio/low_min": 0.000261507185496157, "clip_ratio/region_mean": 0.002125520814054956, "completions/clipped_ratio": 0.0, "completions/max_length": 575.7, "completions/max_terminated_length": 575.7, "completions/mean_length": 198.98333740234375, "completions/mean_terminated_length": 198.98333740234375, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.028432453672091167, "epoch": 23.44566880930517, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3920476952.0, "reward": 8.330292749404908, "reward_std": 0.2949476182460785, "rewards/combined_reward/mean": 8.330292463302612, "rewards/combined_reward/std": 1.8625856757164, "step": 38310 }, { "clip_ratio/high_max": 0.0017789251719174596, "clip_ratio/high_mean": 0.0010357641372441624, "clip_ratio/low_mean": 0.000899392609668818, "clip_ratio/low_min": 0.00025948565453290937, "clip_ratio/region_mean": 0.0019351567641327468, "completions/clipped_ratio": 0.0, "completions/max_length": 543.4, "completions/max_terminated_length": 543.4, "completions/mean_length": 194.85860137939454, "completions/mean_terminated_length": 194.85860137939454, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.027229732655299207, "epoch": 23.4640342822161, "frac_reward_zero_std": 0.15416666865348816, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 3923525369.0, "reward": 8.274580001831055, "reward_std": 0.31555933952331544, "rewards/combined_reward/mean": 8.274579954147338, "rewards/combined_reward/std": 1.9033868670463563, "step": 38340 }, { "clip_ratio/high_max": 0.001974569233910491, "clip_ratio/high_mean": 0.0011462392119331829, "clip_ratio/low_mean": 0.0007484829987030631, "clip_ratio/low_min": 0.00021137067645516557, "clip_ratio/region_mean": 0.0018947222192461292, "completions/clipped_ratio": 0.0, "completions/max_length": 618.6, "completions/max_terminated_length": 618.6, "completions/mean_length": 204.4078170776367, "completions/mean_terminated_length": 204.4078170776367, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.029570010661457976, "epoch": 23.482399755127027, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3926610455.0, "reward": 8.16055736541748, "reward_std": 0.3610701233148575, "rewards/combined_reward/mean": 8.160557174682618, "rewards/combined_reward/std": 1.975184941291809, "step": 38370 }, { "clip_ratio/high_max": 0.002079994119897795, "clip_ratio/high_mean": 0.0012083510449883762, "clip_ratio/low_mean": 0.0008859806079271948, "clip_ratio/low_min": 0.0002346279373644696, "clip_ratio/region_mean": 0.00209433164272923, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 199.48776550292968, "completions/mean_terminated_length": 199.48776550292968, "completions/min_length": 43.4, "completions/min_terminated_length": 43.4, "entropy": 0.03048897817110022, "epoch": 23.500765228037956, "frac_reward_zero_std": 0.17500000223517417, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3929676648.0, "reward": 7.6930163383483885, "reward_std": 0.3212802201509476, "rewards/combined_reward/mean": 7.693016099929809, "rewards/combined_reward/std": 2.535653221607208, "step": 38400 }, { "clip_ratio/high_max": 0.001667335693491623, "clip_ratio/high_mean": 0.001001875693934077, "clip_ratio/low_mean": 0.0007202655582659645, "clip_ratio/low_min": 0.0001909730099820687, "clip_ratio/region_mean": 0.001722141245166616, "completions/clipped_ratio": 0.0, "completions/max_length": 595.3, "completions/max_terminated_length": 595.3, "completions/mean_length": 200.27031707763672, "completions/mean_terminated_length": 200.27031707763672, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.026596143717567126, "epoch": 23.51913070094888, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.0230712890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 3932745846.0, "reward": 7.92526798248291, "reward_std": 0.3206644743680954, "rewards/combined_reward/mean": 7.925267839431763, "rewards/combined_reward/std": 2.349903070926666, "step": 38430 }, { "clip_ratio/high_max": 0.0020977457699094277, "clip_ratio/high_mean": 0.0012201622518963025, "clip_ratio/low_mean": 0.0008130987410064942, "clip_ratio/low_min": 0.00028808580803646087, "clip_ratio/region_mean": 0.0020332609953281158, "completions/clipped_ratio": 0.0, "completions/max_length": 658.7, "completions/max_terminated_length": 658.7, "completions/mean_length": 216.1791732788086, "completions/mean_terminated_length": 216.1791732788086, "completions/min_length": 45.5, "completions/min_terminated_length": 45.5, "entropy": 0.03158432114869356, "epoch": 23.53749617385981, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.026611328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3935876134.0, "reward": 7.685997009277344, "reward_std": 0.3725303441286087, "rewards/combined_reward/mean": 7.685996866226196, "rewards/combined_reward/std": 2.636411964893341, "step": 38460 }, { "clip_ratio/high_max": 0.0017754787467614127, "clip_ratio/high_mean": 0.0010537650712649338, "clip_ratio/low_mean": 0.0007472844773777372, "clip_ratio/low_min": 0.00017839224165072665, "clip_ratio/region_mean": 0.0018010495468236817, "completions/clipped_ratio": 0.0, "completions/max_length": 630.9, "completions/max_terminated_length": 630.9, "completions/mean_length": 198.25469207763672, "completions/mean_terminated_length": 198.25469207763672, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.02779310370484988, "epoch": 23.555861646770737, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3938937592.0, "reward": 8.110668420791626, "reward_std": 0.3400894567370415, "rewards/combined_reward/mean": 8.11066837310791, "rewards/combined_reward/std": 2.185976767539978, "step": 38490 }, { "clip_ratio/high_max": 0.0019816302316030487, "clip_ratio/high_mean": 0.0011706089513609185, "clip_ratio/low_mean": 0.0007301978960943719, "clip_ratio/low_min": 0.000240800017490983, "clip_ratio/region_mean": 0.0019008068586117588, "completions/clipped_ratio": 0.0, "completions/max_length": 500.2, "completions/max_terminated_length": 500.2, "completions/mean_length": 192.39818115234374, "completions/mean_terminated_length": 192.39818115234374, "completions/min_length": 45.5, "completions/min_terminated_length": 45.5, "entropy": 0.028534617833793165, "epoch": 23.574227119681666, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3941976561.0, "reward": 7.982449102401733, "reward_std": 0.3648463174700737, "rewards/combined_reward/mean": 7.982448863983154, "rewards/combined_reward/std": 2.2170071482658384, "step": 38520 }, { "clip_ratio/high_max": 0.0020566432984196583, "clip_ratio/high_mean": 0.0011346151073666987, "clip_ratio/low_mean": 0.0008956636828467405, "clip_ratio/low_min": 0.0002960150435683317, "clip_ratio/region_mean": 0.002030278797610663, "completions/clipped_ratio": 0.0, "completions/max_length": 550.1, "completions/max_terminated_length": 550.1, "completions/mean_length": 196.68750610351563, "completions/mean_terminated_length": 196.68750610351563, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03095569321109603, "epoch": 23.59259259259259, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3945032001.0, "reward": 8.18172836303711, "reward_std": 0.3165434405207634, "rewards/combined_reward/mean": 8.18172812461853, "rewards/combined_reward/std": 2.136116933822632, "step": 38550 }, { "clip_ratio/high_max": 0.0018852478007223303, "clip_ratio/high_mean": 0.0010302044532484917, "clip_ratio/low_mean": 0.0006605770294603037, "clip_ratio/low_min": 0.0001635430230332228, "clip_ratio/region_mean": 0.0016907814894996894, "completions/clipped_ratio": 0.0, "completions/max_length": 545.1, "completions/max_terminated_length": 545.1, "completions/mean_length": 197.40287017822266, "completions/mean_terminated_length": 197.40287017822266, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02910898414750894, "epoch": 23.61095806550352, "frac_reward_zero_std": 0.1791666705161333, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3948090188.0, "reward": 8.205915546417236, "reward_std": 0.316573503613472, "rewards/combined_reward/mean": 8.205915307998657, "rewards/combined_reward/std": 2.224585974216461, "step": 38580 }, { "clip_ratio/high_max": 0.0020237930545893806, "clip_ratio/high_mean": 0.0011463958066694128, "clip_ratio/low_mean": 0.0007372973625630645, "clip_ratio/low_min": 0.00016966372277238406, "clip_ratio/region_mean": 0.0018836931665039932, "completions/clipped_ratio": 0.0, "completions/max_length": 641.3, "completions/max_terminated_length": 641.3, "completions/mean_length": 195.45209045410155, "completions/mean_terminated_length": 195.45209045410155, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028687012086932858, "epoch": 23.629323538414447, "frac_reward_zero_std": 0.18750000707805156, "grad_norm": 0.0361328125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 3951140884.0, "reward": 7.931820440292358, "reward_std": 0.33223944902420044, "rewards/combined_reward/mean": 7.931820249557495, "rewards/combined_reward/std": 2.5531253814697266, "step": 38610 }, { "clip_ratio/high_max": 0.0018126797459747954, "clip_ratio/high_mean": 0.0010869894878851483, "clip_ratio/low_mean": 0.0007664806857064832, "clip_ratio/low_min": 0.00020762634934120192, "clip_ratio/region_mean": 0.0018534701773508762, "completions/clipped_ratio": 0.0, "completions/max_length": 525.8, "completions/max_terminated_length": 525.8, "completions/mean_length": 187.24713897705078, "completions/mean_terminated_length": 187.24713897705078, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.028754082787781955, "epoch": 23.647689011325376, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3954160073.0, "reward": 8.139653491973878, "reward_std": 0.3201427534222603, "rewards/combined_reward/mean": 8.13965311050415, "rewards/combined_reward/std": 2.107051360607147, "step": 38640 }, { "clip_ratio/high_max": 0.0017386136707500555, "clip_ratio/high_mean": 0.0010541442137764535, "clip_ratio/low_mean": 0.0008237375146563864, "clip_ratio/low_min": 0.00023238758391623074, "clip_ratio/region_mean": 0.0018778817534136276, "completions/clipped_ratio": 0.0, "completions/max_length": 547.5, "completions/max_terminated_length": 547.5, "completions/mean_length": 211.69662170410157, "completions/mean_terminated_length": 211.69662170410157, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.028656427220751842, "epoch": 23.6660544842363, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 3957273148.0, "reward": 7.793717765808106, "reward_std": 0.459457078576088, "rewards/combined_reward/mean": 7.793717718124389, "rewards/combined_reward/std": 2.488979697227478, "step": 38670 }, { "clip_ratio/high_max": 0.0019438215822447091, "clip_ratio/high_mean": 0.0010543964024691376, "clip_ratio/low_mean": 0.0008341233601337687, "clip_ratio/low_min": 0.00020933291865124677, "clip_ratio/region_mean": 0.0018885197593287255, "completions/clipped_ratio": 0.0, "completions/max_length": 626.5, "completions/max_terminated_length": 626.5, "completions/mean_length": 189.8492218017578, "completions/mean_terminated_length": 189.8492218017578, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028734260727651416, "epoch": 23.68441995714723, "frac_reward_zero_std": 0.1625000063329935, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3960302329.0, "reward": 7.84889783859253, "reward_std": 0.2974908113479614, "rewards/combined_reward/mean": 7.8488976001739506, "rewards/combined_reward/std": 2.283720552921295, "step": 38700 }, { "clip_ratio/high_max": 0.002115017277537845, "clip_ratio/high_mean": 0.0012461132822257544, "clip_ratio/low_mean": 0.0007597836307468242, "clip_ratio/low_min": 0.0001549170723592397, "clip_ratio/region_mean": 0.0020058969224919566, "completions/clipped_ratio": 0.0, "completions/max_length": 578.8, "completions/max_terminated_length": 578.8, "completions/mean_length": 190.3804733276367, "completions/mean_terminated_length": 190.3804733276367, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028773168140711885, "epoch": 23.702785430058157, "frac_reward_zero_std": 0.20416667684912682, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 3963333550.0, "reward": 7.541000890731811, "reward_std": 0.34882189333438873, "rewards/combined_reward/mean": 7.541000747680664, "rewards/combined_reward/std": 2.699190640449524, "step": 38730 }, { "clip_ratio/high_max": 0.0017380991014457928, "clip_ratio/high_mean": 0.0010244825372258978, "clip_ratio/low_mean": 0.0007629425318251985, "clip_ratio/low_min": 0.00022379209840437398, "clip_ratio/region_mean": 0.0017874250717189474, "completions/clipped_ratio": 0.0, "completions/max_length": 544.4, "completions/max_terminated_length": 544.4, "completions/mean_length": 201.14844207763673, "completions/mean_terminated_length": 201.14844207763673, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.02643537463930746, "epoch": 23.721150902969086, "frac_reward_zero_std": 0.14583333693444728, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3966406120.0, "reward": 8.10149164199829, "reward_std": 0.2960920289158821, "rewards/combined_reward/mean": 8.101491117477417, "rewards/combined_reward/std": 2.175101912021637, "step": 38760 }, { "clip_ratio/high_max": 0.0019331992143027795, "clip_ratio/high_mean": 0.0010632248252174274, "clip_ratio/low_mean": 0.0007308865143083192, "clip_ratio/low_min": 0.00016270586541698625, "clip_ratio/region_mean": 0.0017941113066626713, "completions/clipped_ratio": 0.0, "completions/max_length": 567.8, "completions/max_terminated_length": 567.8, "completions/mean_length": 198.12839050292968, "completions/mean_terminated_length": 198.12839050292968, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.02803075159123788, "epoch": 23.73951637588001, "frac_reward_zero_std": 0.20833333507180213, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 3969467093.0, "reward": 8.028707551956177, "reward_std": 0.3204744875431061, "rewards/combined_reward/mean": 8.028707361221313, "rewards/combined_reward/std": 2.465301585197449, "step": 38790 }, { "clip_ratio/high_max": 0.0020504944055574014, "clip_ratio/high_mean": 0.00125387697359353, "clip_ratio/low_mean": 0.0007024831469607306, "clip_ratio/low_min": 0.00021632504843485851, "clip_ratio/region_mean": 0.001956360134257314, "completions/clipped_ratio": 0.0, "completions/max_length": 618.5, "completions/max_terminated_length": 618.5, "completions/mean_length": 210.31380767822264, "completions/mean_terminated_length": 210.31380767822264, "completions/min_length": 51.7, "completions/min_terminated_length": 51.7, "entropy": 0.03154275399477532, "epoch": 23.75788184879094, "frac_reward_zero_std": 0.12916667014360428, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 3972574858.0, "reward": 7.9997282981872555, "reward_std": 0.35769837647676467, "rewards/combined_reward/mean": 7.999728059768676, "rewards/combined_reward/std": 2.38831250667572, "step": 38820 }, { "clip_ratio/high_max": 0.001522971742815571, "clip_ratio/high_mean": 0.0008888636505920052, "clip_ratio/low_mean": 0.0006227006818032047, "clip_ratio/low_min": 0.00019310539913324948, "clip_ratio/region_mean": 0.0015115643237853269, "completions/clipped_ratio": 0.0, "completions/max_length": 627.7, "completions/max_terminated_length": 627.7, "completions/mean_length": 194.23099517822266, "completions/mean_terminated_length": 194.23099517822266, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.023675425684389968, "epoch": 23.776247321701867, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 3975620865.0, "reward": 8.412667274475098, "reward_std": 0.26342013776302337, "rewards/combined_reward/mean": 8.412667083740235, "rewards/combined_reward/std": 1.8822748839855195, "step": 38850 }, { "clip_ratio/high_max": 0.001891747447856081, "clip_ratio/high_mean": 0.0011391342406568584, "clip_ratio/low_mean": 0.0006688036200406107, "clip_ratio/low_min": 0.0002117375704983715, "clip_ratio/region_mean": 0.001807937851117458, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 208.10599517822266, "completions/mean_terminated_length": 208.10599517822266, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.02760703794968625, "epoch": 23.794612794612796, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.0150146484375, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 3978720152.0, "reward": 8.083470010757447, "reward_std": 0.32006202936172484, "rewards/combined_reward/mean": 8.083469820022582, "rewards/combined_reward/std": 2.300072729587555, "step": 38880 }, { "clip_ratio/high_max": 0.0024769547293544748, "clip_ratio/high_mean": 0.001415917959335881, "clip_ratio/low_mean": 0.00097543372491297, "clip_ratio/low_min": 0.0002483680410174808, "clip_ratio/region_mean": 0.002391351708016979, "completions/clipped_ratio": 0.0, "completions/max_length": 689.6, "completions/max_terminated_length": 689.6, "completions/mean_length": 204.04896392822266, "completions/mean_terminated_length": 204.04896392822266, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.03359976878079275, "epoch": 23.81297826752372, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.025146484375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 3981803860.0, "reward": 7.9843230724334715, "reward_std": 0.33153113573789594, "rewards/combined_reward/mean": 7.9843230724334715, "rewards/combined_reward/std": 2.1807436466217043, "step": 38910 }, { "clip_ratio/high_max": 0.0020390567272746316, "clip_ratio/high_mean": 0.0012346770050498889, "clip_ratio/low_mean": 0.0007900977208919358, "clip_ratio/low_min": 0.0002542817737169874, "clip_ratio/region_mean": 0.0020247747292160057, "completions/clipped_ratio": 0.0, "completions/max_length": 605.9, "completions/max_terminated_length": 605.9, "completions/mean_length": 196.2609436035156, "completions/mean_terminated_length": 196.2609436035156, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.031020641761521497, "epoch": 23.83134374043465, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.006927490234375, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 3984857662.0, "reward": 7.794880485534668, "reward_std": 0.34317106902599337, "rewards/combined_reward/mean": 7.794880342483521, "rewards/combined_reward/std": 2.506417715549469, "step": 38940 }, { "clip_ratio/high_max": 0.0016750062764913308, "clip_ratio/high_mean": 0.0009323753771847502, "clip_ratio/low_mean": 0.0006810351801201857, "clip_ratio/low_min": 0.00018304957629879936, "clip_ratio/region_mean": 0.0016134105499077123, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 193.4289108276367, "completions/mean_terminated_length": 193.4289108276367, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.025098982996617756, "epoch": 23.849709213345577, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 3987900589.0, "reward": 8.186307907104492, "reward_std": 0.2664533004164696, "rewards/combined_reward/mean": 8.186307621002197, "rewards/combined_reward/std": 2.2196478486061095, "step": 38970 }, { "clip_ratio/high_max": 0.0018956086506174566, "clip_ratio/high_mean": 0.0010993946578916316, "clip_ratio/low_mean": 0.0006532083731144667, "clip_ratio/low_min": 0.00019816001367871648, "clip_ratio/region_mean": 0.0017526030404648432, "completions/clipped_ratio": 0.0, "completions/max_length": 605.7, "completions/max_terminated_length": 605.7, "completions/mean_length": 197.82891235351562, "completions/mean_terminated_length": 197.82891235351562, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.025113419629633427, "epoch": 23.868074686256506, "frac_reward_zero_std": 0.16666667126119136, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 3990960412.0, "reward": 8.0387291431427, "reward_std": 0.3028073519468307, "rewards/combined_reward/mean": 8.038728761672974, "rewards/combined_reward/std": 2.428061544895172, "step": 39000 }, { "clip_ratio/high_max": 0.0021305421784442538, "clip_ratio/high_mean": 0.0012564204596856143, "clip_ratio/low_mean": 0.0007968078606306032, "clip_ratio/low_min": 0.0002507956681559638, "clip_ratio/region_mean": 0.002053228343720548, "completions/clipped_ratio": 0.0, "completions/max_length": 597.3, "completions/max_terminated_length": 597.3, "completions/mean_length": 210.19688110351564, "completions/mean_terminated_length": 210.19688110351564, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03284212929817538, "epoch": 23.88644015916743, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 3994067728.0, "reward": 7.992381954193116, "reward_std": 0.4205131411552429, "rewards/combined_reward/mean": 7.992381954193116, "rewards/combined_reward/std": 2.228304851055145, "step": 39030 }, { "clip_ratio/high_max": 0.001788447850412922, "clip_ratio/high_mean": 0.001037466681736987, "clip_ratio/low_mean": 0.0007629194120454486, "clip_ratio/low_min": 0.00019587989518186078, "clip_ratio/region_mean": 0.0018003861073035902, "completions/clipped_ratio": 0.0, "completions/max_length": 577.2, "completions/max_terminated_length": 577.2, "completions/mean_length": 203.06198425292968, "completions/mean_terminated_length": 203.06198425292968, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02944102257800599, "epoch": 23.90480563207836, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.0277099609375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 3997147646.0, "reward": 8.275331354141235, "reward_std": 0.2966182127594948, "rewards/combined_reward/mean": 8.275331258773804, "rewards/combined_reward/std": 1.9979158639907837, "step": 39060 }, { "clip_ratio/high_max": 0.002182170783150165, "clip_ratio/high_mean": 0.001209993036900414, "clip_ratio/low_mean": 0.0007444726873776139, "clip_ratio/low_min": 0.0001886904605877741, "clip_ratio/region_mean": 0.0019544657217920756, "completions/clipped_ratio": 0.0, "completions/max_length": 612.7, "completions/max_terminated_length": 612.7, "completions/mean_length": 209.80859832763673, "completions/mean_terminated_length": 209.80859832763673, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03194271173949043, "epoch": 23.923171104989287, "frac_reward_zero_std": 0.11666666828095913, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4000253471.0, "reward": 7.754737091064453, "reward_std": 0.3403680294752121, "rewards/combined_reward/mean": 7.754736757278442, "rewards/combined_reward/std": 2.413537383079529, "step": 39090 }, { "clip_ratio/high_max": 0.001803585018690986, "clip_ratio/high_mean": 0.0009931316356717919, "clip_ratio/low_mean": 0.0007212499131128425, "clip_ratio/low_min": 0.00016380132268144129, "clip_ratio/region_mean": 0.0017143815384770278, "completions/clipped_ratio": 0.0, "completions/max_length": 521.5, "completions/max_terminated_length": 521.5, "completions/mean_length": 191.63516082763672, "completions/mean_terminated_length": 191.63516082763672, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.026480041320125262, "epoch": 23.941536577900216, "frac_reward_zero_std": 0.20833334177732468, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4003289510.0, "reward": 8.112214183807373, "reward_std": 0.2796778783202171, "rewards/combined_reward/mean": 8.1122136592865, "rewards/combined_reward/std": 1.9636057376861573, "step": 39120 }, { "clip_ratio/high_max": 0.002014268181907634, "clip_ratio/high_mean": 0.0011703186736364537, "clip_ratio/low_mean": 0.0006332202210614923, "clip_ratio/low_min": 0.00019267565270032113, "clip_ratio/region_mean": 0.0018035388881495844, "completions/clipped_ratio": 0.0, "completions/max_length": 572.1, "completions/max_terminated_length": 572.1, "completions/mean_length": 215.7096405029297, "completions/mean_terminated_length": 215.7096405029297, "completions/min_length": 49.8, "completions/min_terminated_length": 49.8, "entropy": 0.027254372540240487, "epoch": 23.95990205081114, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.004974365234375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4006417995.0, "reward": 7.981906175613403, "reward_std": 0.35004455149173735, "rewards/combined_reward/mean": 7.981905746459961, "rewards/combined_reward/std": 2.3556735634803774, "step": 39150 }, { "clip_ratio/high_max": 0.0016066367146171008, "clip_ratio/high_mean": 0.0008973639487521723, "clip_ratio/low_mean": 0.0008011342146346578, "clip_ratio/low_min": 0.00025926515566728386, "clip_ratio/region_mean": 0.001698498166903543, "completions/clipped_ratio": 0.0, "completions/max_length": 602.5, "completions/max_terminated_length": 602.5, "completions/mean_length": 200.2687530517578, "completions/mean_terminated_length": 200.2687530517578, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028017681650817396, "epoch": 23.97826752372207, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4009487187.0, "reward": 8.335190916061402, "reward_std": 0.32338821440935134, "rewards/combined_reward/mean": 8.335190534591675, "rewards/combined_reward/std": 1.8852523803710937, "step": 39180 }, { "clip_ratio/high_max": 0.001894432735571172, "clip_ratio/high_mean": 0.001009441225566358, "clip_ratio/low_mean": 0.0007349146607642372, "clip_ratio/low_min": 0.00018096426744402077, "clip_ratio/region_mean": 0.001744355868625765, "completions/clipped_ratio": 0.0, "completions/max_length": 625.4, "completions/max_terminated_length": 625.4, "completions/mean_length": 199.65625457763673, "completions/mean_terminated_length": 199.65625457763673, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.026863213892405233, "epoch": 23.996632996632997, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4012554027.0, "reward": 8.008767414093018, "reward_std": 0.28751862943172457, "rewards/combined_reward/mean": 8.00876727104187, "rewards/combined_reward/std": 2.2816386461257934, "step": 39210 }, { "clip_ratio/high_max": 0.0023621174365617945, "clip_ratio/high_mean": 0.00128669895663033, "clip_ratio/low_mean": 0.0008370499486832681, "clip_ratio/low_min": 0.0002324547020806244, "clip_ratio/region_mean": 0.0021237489198655136, "completions/clipped_ratio": 0.0, "completions/max_length": 602.9, "completions/max_terminated_length": 602.9, "completions/mean_length": 206.8947982788086, "completions/mean_terminated_length": 206.8947982788086, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026853175116244885, "epoch": 24.015304560759105, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.01116943359375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4015648663.0, "reward": 8.113206768035889, "reward_std": 0.3260280877351761, "rewards/combined_reward/mean": 8.113206100463866, "rewards/combined_reward/std": 2.198161554336548, "step": 39240 }, { "clip_ratio/high_max": 0.0023375718051587076, "clip_ratio/high_mean": 0.0014127287090862713, "clip_ratio/low_mean": 0.0009481506718050999, "clip_ratio/low_min": 0.000300153456676829, "clip_ratio/region_mean": 0.0023608793802850415, "completions/clipped_ratio": 0.0, "completions/max_length": 665.1, "completions/max_terminated_length": 665.1, "completions/mean_length": 205.1867248535156, "completions/mean_terminated_length": 205.1867248535156, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.02941337033795814, "epoch": 24.033670033670035, "frac_reward_zero_std": 0.1250000014901161, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4018736740.0, "reward": 8.219331359863281, "reward_std": 0.3806120574474335, "rewards/combined_reward/mean": 8.21933102607727, "rewards/combined_reward/std": 2.136263883113861, "step": 39270 }, { "clip_ratio/high_max": 0.0023659377914251917, "clip_ratio/high_mean": 0.0012632000593536456, "clip_ratio/low_mean": 0.0007150265212355104, "clip_ratio/low_min": 0.00021273508076167976, "clip_ratio/region_mean": 0.001978226583499539, "completions/clipped_ratio": 0.0, "completions/max_length": 542.7, "completions/max_terminated_length": 542.7, "completions/mean_length": 199.7338607788086, "completions/mean_terminated_length": 199.7338607788086, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.026395448002343377, "epoch": 24.05203550658096, "frac_reward_zero_std": 0.15833333916962147, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4021803878.0, "reward": 7.928727769851685, "reward_std": 0.4004669576883316, "rewards/combined_reward/mean": 7.928727817535401, "rewards/combined_reward/std": 2.3563855409622194, "step": 39300 }, { "clip_ratio/high_max": 0.0022864449446691045, "clip_ratio/high_mean": 0.0012179291410575389, "clip_ratio/low_mean": 0.0009131037585272376, "clip_ratio/low_min": 0.0002553531667217612, "clip_ratio/region_mean": 0.002131032895219202, "completions/clipped_ratio": 0.0, "completions/max_length": 526.9, "completions/max_terminated_length": 526.9, "completions/mean_length": 196.32161865234374, "completions/mean_terminated_length": 196.32161865234374, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02730420557782054, "epoch": 24.07040097949189, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.005096435546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4024857913.0, "reward": 8.182015132904052, "reward_std": 0.2620679810643196, "rewards/combined_reward/mean": 8.182014894485473, "rewards/combined_reward/std": 2.163823962211609, "step": 39330 }, { "clip_ratio/high_max": 0.002385718885489041, "clip_ratio/high_mean": 0.0013108110455505085, "clip_ratio/low_mean": 0.0008583875624026405, "clip_ratio/low_min": 0.00021935005812944534, "clip_ratio/region_mean": 0.0021691986029812446, "completions/clipped_ratio": 0.0, "completions/max_length": 543.9, "completions/max_terminated_length": 543.9, "completions/mean_length": 208.22891082763672, "completions/mean_terminated_length": 208.22891082763672, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026295277414222558, "epoch": 24.088766452402815, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4027957672.0, "reward": 8.104894113540649, "reward_std": 0.34089784026145936, "rewards/combined_reward/mean": 8.104894161224365, "rewards/combined_reward/std": 2.263654553890228, "step": 39360 }, { "clip_ratio/high_max": 0.0022841228725155815, "clip_ratio/high_mean": 0.0012937808501495359, "clip_ratio/low_mean": 0.0008704323608981213, "clip_ratio/low_min": 0.0002731216183747165, "clip_ratio/region_mean": 0.0021642132148069018, "completions/clipped_ratio": 0.0, "completions/max_length": 721.4, "completions/max_terminated_length": 721.4, "completions/mean_length": 213.3546920776367, "completions/mean_terminated_length": 213.3546920776367, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02898138143743078, "epoch": 24.107131925313745, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4031077114.0, "reward": 8.151456165313721, "reward_std": 0.34517953246831895, "rewards/combined_reward/mean": 8.15145583152771, "rewards/combined_reward/std": 2.223524343967438, "step": 39390 }, { "clip_ratio/high_max": 0.002565202108720162, "clip_ratio/high_mean": 0.00150453877261801, "clip_ratio/low_mean": 0.0009579673403398677, "clip_ratio/low_min": 0.00024454503024268584, "clip_ratio/region_mean": 0.0024625061225378884, "completions/clipped_ratio": 0.0, "completions/max_length": 562.2, "completions/max_terminated_length": 562.2, "completions/mean_length": 203.26068267822265, "completions/mean_terminated_length": 203.26068267822265, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.03151485666166991, "epoch": 24.12549739822467, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4034157795.0, "reward": 8.172011184692384, "reward_std": 0.332199402153492, "rewards/combined_reward/mean": 8.172010898590088, "rewards/combined_reward/std": 2.103078281879425, "step": 39420 }, { "clip_ratio/high_max": 0.0022064531100719855, "clip_ratio/high_mean": 0.0013061449153004409, "clip_ratio/low_mean": 0.0009096434189511152, "clip_ratio/low_min": 0.0002658128219385011, "clip_ratio/region_mean": 0.002215788341224349, "completions/clipped_ratio": 0.0, "completions/max_length": 574.2, "completions/max_terminated_length": 574.2, "completions/mean_length": 203.48438110351563, "completions/mean_terminated_length": 203.48438110351563, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.028818328919199606, "epoch": 24.1438628711356, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.0311279296875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4037239335.0, "reward": 8.00252652168274, "reward_std": 0.3593684405088425, "rewards/combined_reward/mean": 8.002526473999023, "rewards/combined_reward/std": 2.3013013124465944, "step": 39450 }, { "clip_ratio/high_max": 0.0027583214522261795, "clip_ratio/high_mean": 0.0015255516775747916, "clip_ratio/low_mean": 0.000877446414839748, "clip_ratio/low_min": 0.0002473176352699132, "clip_ratio/region_mean": 0.002402998089382891, "completions/clipped_ratio": 0.0, "completions/max_length": 581.2, "completions/max_terminated_length": 581.2, "completions/mean_length": 207.69219207763672, "completions/mean_terminated_length": 207.69219207763672, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02927585649304092, "epoch": 24.162228344046525, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.008056640625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4040337033.0, "reward": 8.310390996932984, "reward_std": 0.30012809187173844, "rewards/combined_reward/mean": 8.310390710830688, "rewards/combined_reward/std": 1.9940425992012023, "step": 39480 }, { "clip_ratio/high_max": 0.0023354883895081004, "clip_ratio/high_mean": 0.0012911466870718868, "clip_ratio/low_mean": 0.0009391062997262149, "clip_ratio/low_min": 0.0003108940589299891, "clip_ratio/region_mean": 0.002230252985221644, "completions/clipped_ratio": 0.0, "completions/max_length": 599.8, "completions/max_terminated_length": 599.8, "completions/mean_length": 205.81250457763673, "completions/mean_terminated_length": 205.81250457763673, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.027947557407120863, "epoch": 24.180593816957455, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4043427513.0, "reward": 8.174535942077636, "reward_std": 0.28374144732952117, "rewards/combined_reward/mean": 8.174535655975342, "rewards/combined_reward/std": 2.1684762477874755, "step": 39510 }, { "clip_ratio/high_max": 0.0033569185024437806, "clip_ratio/high_mean": 0.0018436336573358857, "clip_ratio/low_mean": 0.000906116916545822, "clip_ratio/low_min": 0.00015784727099041145, "clip_ratio/region_mean": 0.002749750578853612, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 186.89453887939453, "completions/mean_terminated_length": 186.89453887939453, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.032466311908016604, "epoch": 24.19895928986838, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4046445348.0, "reward": 8.179482030868531, "reward_std": 0.3491471394896507, "rewards/combined_reward/mean": 8.179481649398804, "rewards/combined_reward/std": 2.2117191791534423, "step": 39540 }, { "clip_ratio/high_max": 0.002140774245102269, "clip_ratio/high_mean": 0.0012136270499468083, "clip_ratio/low_mean": 0.00088410994176229, "clip_ratio/low_min": 0.00026103061876104526, "clip_ratio/region_mean": 0.002097737001652907, "completions/clipped_ratio": 0.0, "completions/max_length": 680.8, "completions/max_terminated_length": 680.8, "completions/mean_length": 201.8197982788086, "completions/mean_terminated_length": 201.8197982788086, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.025238321030822894, "epoch": 24.21732476277931, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4049520496.0, "reward": 8.121022176742553, "reward_std": 0.2971134051680565, "rewards/combined_reward/mean": 8.121022033691407, "rewards/combined_reward/std": 2.2158933162689207, "step": 39570 }, { "clip_ratio/high_max": 0.002106354306791521, "clip_ratio/high_mean": 0.0012357112947938731, "clip_ratio/low_mean": 0.000953947649274293, "clip_ratio/low_min": 0.00021833773207617923, "clip_ratio/region_mean": 0.002189658956922358, "completions/clipped_ratio": 0.0, "completions/max_length": 568.2, "completions/max_terminated_length": 568.2, "completions/mean_length": 197.3987030029297, "completions/mean_terminated_length": 197.3987030029297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027038105810061098, "epoch": 24.235690235690235, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4052578667.0, "reward": 7.96161732673645, "reward_std": 0.3289435088634491, "rewards/combined_reward/mean": 7.961617040634155, "rewards/combined_reward/std": 2.3527451515197755, "step": 39600 }, { "clip_ratio/high_max": 0.0024615995076601393, "clip_ratio/high_mean": 0.001345885931975014, "clip_ratio/low_mean": 0.0008110172862264638, "clip_ratio/low_min": 0.00017830585905661186, "clip_ratio/region_mean": 0.0021569032251136377, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 197.6276092529297, "completions/mean_terminated_length": 197.6276092529297, "completions/min_length": 47.5, "completions/min_terminated_length": 47.5, "entropy": 0.027857389332105718, "epoch": 24.254055708601165, "frac_reward_zero_std": 0.15416667312383653, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4055637717.0, "reward": 8.092267894744873, "reward_std": 0.3188485026359558, "rewards/combined_reward/mean": 8.092267560958863, "rewards/combined_reward/std": 2.1555312156677244, "step": 39630 }, { "clip_ratio/high_max": 0.002963302448915783, "clip_ratio/high_mean": 0.0016833642228448297, "clip_ratio/low_mean": 0.0008132805316563463, "clip_ratio/low_min": 0.0002365814723210254, "clip_ratio/region_mean": 0.002496644767234102, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 204.21563262939452, "completions/mean_terminated_length": 204.21563262939452, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.033477522308627765, "epoch": 24.27242118151209, "frac_reward_zero_std": 0.17500000149011613, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4058722065.0, "reward": 8.243773412704467, "reward_std": 0.38843475878238676, "rewards/combined_reward/mean": 8.243773221969604, "rewards/combined_reward/std": 1.9818347096443176, "step": 39660 }, { "clip_ratio/high_max": 0.002412476332877607, "clip_ratio/high_mean": 0.0013260336522459208, "clip_ratio/low_mean": 0.0009093035051288705, "clip_ratio/low_min": 0.0002669733864119432, "clip_ratio/region_mean": 0.0022353371711991104, "completions/clipped_ratio": 0.0, "completions/max_length": 593.5, "completions/max_terminated_length": 593.5, "completions/mean_length": 192.2609420776367, "completions/mean_terminated_length": 192.2609420776367, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027400043280795216, "epoch": 24.29078665442302, "frac_reward_zero_std": 0.21666667610406876, "grad_norm": 0.01434326171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4061760507.0, "reward": 7.90727391242981, "reward_std": 0.2984308511018753, "rewards/combined_reward/mean": 7.907273769378662, "rewards/combined_reward/std": 2.4261614203453066, "step": 39690 }, { "clip_ratio/high_max": 0.0028404504351783543, "clip_ratio/high_mean": 0.0015555351361399516, "clip_ratio/low_mean": 0.0008212705990445102, "clip_ratio/low_min": 0.0001333244009098659, "clip_ratio/region_mean": 0.002376805713962919, "completions/clipped_ratio": 0.0, "completions/max_length": 542.9, "completions/max_terminated_length": 542.9, "completions/mean_length": 196.40443267822266, "completions/mean_terminated_length": 196.40443267822266, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.03143203981841604, "epoch": 24.309152127333945, "frac_reward_zero_std": 0.18750000409781933, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4064814860.0, "reward": 8.134224605560302, "reward_std": 0.350632943212986, "rewards/combined_reward/mean": 8.13422441482544, "rewards/combined_reward/std": 2.2502902030944822, "step": 39720 }, { "clip_ratio/high_max": 0.0025599966111864584, "clip_ratio/high_mean": 0.0014171939030347858, "clip_ratio/low_mean": 0.0010107473171350042, "clip_ratio/low_min": 0.00028482547180222657, "clip_ratio/region_mean": 0.002427941223140806, "completions/clipped_ratio": 0.0, "completions/max_length": 652.9, "completions/max_terminated_length": 652.9, "completions/mean_length": 197.00469360351562, "completions/mean_terminated_length": 197.00469360351562, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03282678732648492, "epoch": 24.327517600244875, "frac_reward_zero_std": 0.15000000260770321, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4067871518.0, "reward": 7.952191638946533, "reward_std": 0.29293320029973985, "rewards/combined_reward/mean": 7.952191591262817, "rewards/combined_reward/std": 2.382015883922577, "step": 39750 }, { "clip_ratio/high_max": 0.0023895723832538353, "clip_ratio/high_mean": 0.0013732031865705116, "clip_ratio/low_mean": 0.0008816742123599397, "clip_ratio/low_min": 0.0002107498296633518, "clip_ratio/region_mean": 0.0022548774043874196, "completions/clipped_ratio": 0.0, "completions/max_length": 582.6, "completions/max_terminated_length": 582.6, "completions/mean_length": 190.59557647705077, "completions/mean_terminated_length": 190.59557647705077, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.025701801688410342, "epoch": 24.3458830731558, "frac_reward_zero_std": 0.18333333916962147, "grad_norm": 0.0311279296875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4070903565.0, "reward": 8.259976053237915, "reward_std": 0.2648881793022156, "rewards/combined_reward/mean": 8.25997543334961, "rewards/combined_reward/std": 2.0435487389564515, "step": 39780 }, { "clip_ratio/high_max": 0.0024768787562303865, "clip_ratio/high_mean": 0.0013489921084935, "clip_ratio/low_mean": 0.0008123583731503459, "clip_ratio/low_min": 0.00021450362255563958, "clip_ratio/region_mean": 0.0021613504975296865, "completions/clipped_ratio": 0.0, "completions/max_length": 654.1, "completions/max_terminated_length": 654.1, "completions/mean_length": 202.86953887939453, "completions/mean_terminated_length": 202.86953887939453, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02790623117859165, "epoch": 24.36424854606673, "frac_reward_zero_std": 0.14583333693444728, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4073982744.0, "reward": 8.278131914138793, "reward_std": 0.30778928846120834, "rewards/combined_reward/mean": 8.2781316280365, "rewards/combined_reward/std": 2.0078516840934753, "step": 39810 }, { "clip_ratio/high_max": 0.0026296308802557177, "clip_ratio/high_mean": 0.0015641973475188327, "clip_ratio/low_mean": 0.0009827983353413098, "clip_ratio/low_min": 0.0002955925784287198, "clip_ratio/region_mean": 0.0025469956618811313, "completions/clipped_ratio": 0.0, "completions/max_length": 546.9, "completions/max_terminated_length": 546.9, "completions/mean_length": 198.6010482788086, "completions/mean_terminated_length": 198.6010482788086, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03189984407896797, "epoch": 24.382614018977655, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.0247802734375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4077045532.0, "reward": 8.212778091430664, "reward_std": 0.31380238980054853, "rewards/combined_reward/mean": 8.212778186798095, "rewards/combined_reward/std": 2.1181939125061033, "step": 39840 }, { "clip_ratio/high_max": 0.0023245506051656167, "clip_ratio/high_mean": 0.001304984746335928, "clip_ratio/low_mean": 0.000871490603943433, "clip_ratio/low_min": 0.00018004787101138692, "clip_ratio/region_mean": 0.0021764753570702546, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 183.71953582763672, "completions/mean_terminated_length": 183.71953582763672, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "entropy": 0.02625270189406971, "epoch": 24.400979491888585, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.005096435546875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4080051175.0, "reward": 8.339767408370971, "reward_std": 0.2321361631155014, "rewards/combined_reward/mean": 8.339767169952392, "rewards/combined_reward/std": 1.9426934242248535, "step": 39870 }, { "clip_ratio/high_max": 0.0023264164008044946, "clip_ratio/high_mean": 0.0013177783987581885, "clip_ratio/low_mean": 0.0008779952794914909, "clip_ratio/low_min": 0.00022801148661528715, "clip_ratio/region_mean": 0.0021957736935291905, "completions/clipped_ratio": 0.0, "completions/max_length": 704.3, "completions/max_terminated_length": 704.3, "completions/mean_length": 202.33438110351562, "completions/mean_terminated_length": 202.33438110351562, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.026997256302274765, "epoch": 24.41934496479951, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.005706787109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4083128299.0, "reward": 7.980271291732788, "reward_std": 0.31278268843889234, "rewards/combined_reward/mean": 7.980270767211914, "rewards/combined_reward/std": 2.268389892578125, "step": 39900 }, { "clip_ratio/high_max": 0.002328276346088387, "clip_ratio/high_mean": 0.0012857411961401037, "clip_ratio/low_mean": 0.0009422851865868628, "clip_ratio/low_min": 0.00026023367948558494, "clip_ratio/region_mean": 0.002228026378240126, "completions/clipped_ratio": 0.0, "completions/max_length": 595.4, "completions/max_terminated_length": 595.4, "completions/mean_length": 190.57969360351564, "completions/mean_terminated_length": 190.57969360351564, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028925357898697258, "epoch": 24.437710437710436, "frac_reward_zero_std": 0.18750000707805156, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4086160285.0, "reward": 7.9528007984161375, "reward_std": 0.30395677536725996, "rewards/combined_reward/mean": 7.952800464630127, "rewards/combined_reward/std": 2.5338857412338256, "step": 39930 }, { "clip_ratio/high_max": 0.0026742793818508898, "clip_ratio/high_mean": 0.0016490873904937567, "clip_ratio/low_mean": 0.0009948366953418978, "clip_ratio/low_min": 0.0002546801600450029, "clip_ratio/region_mean": 0.0026439240978409846, "completions/clipped_ratio": 0.0, "completions/max_length": 544.2, "completions/max_terminated_length": 544.2, "completions/mean_length": 198.9588607788086, "completions/mean_terminated_length": 198.9588607788086, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.031755539728328584, "epoch": 24.456075910621365, "frac_reward_zero_std": 0.1583333395421505, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4089224447.0, "reward": 7.916634464263916, "reward_std": 0.41209791600704193, "rewards/combined_reward/mean": 7.916634035110474, "rewards/combined_reward/std": 2.4450852274894714, "step": 39960 }, { "clip_ratio/high_max": 0.0021633675540215338, "clip_ratio/high_mean": 0.0011749617270349214, "clip_ratio/low_mean": 0.000976914892210819, "clip_ratio/low_min": 0.00022226438450161367, "clip_ratio/region_mean": 0.002151876611969783, "completions/clipped_ratio": 0.0, "completions/max_length": 624.7, "completions/max_terminated_length": 624.7, "completions/mean_length": 201.97552490234375, "completions/mean_terminated_length": 201.97552490234375, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.026132600397492448, "epoch": 24.47444138353229, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4092300193.0, "reward": 8.15729193687439, "reward_std": 0.25731446146965026, "rewards/combined_reward/mean": 8.157291650772095, "rewards/combined_reward/std": 2.2571695923805235, "step": 39990 }, { "clip_ratio/high_max": 0.001788954623771133, "clip_ratio/high_mean": 0.001014882282106555, "clip_ratio/low_mean": 0.0006758703395462363, "clip_ratio/low_min": 0.00020558232463372406, "clip_ratio/region_mean": 0.0016907526340219192, "completions/clipped_ratio": 0.0, "completions/max_length": 533.4285714285714, "completions/max_terminated_length": 533.4285714285714, "completions/mean_length": 204.63802664620536, "completions/mean_terminated_length": 204.63802664620536, "completions/min_length": 48.42857142857143, "completions/min_terminated_length": 48.42857142857143, "entropy": 0.02727956548333168, "epoch": 24.492194674012858, "frac_reward_zero_std": 0.16666666950498307, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4095371841.0, "reward": 7.910318715231759, "reward_std": 0.39659594212259563, "rewards/combined_reward/mean": 7.910318715231759, "rewards/combined_reward/std": 2.205631307193211, "step": 40020 }, { "clip_ratio/high_max": 0.0018259544136526528, "clip_ratio/high_mean": 0.0010804935936903348, "clip_ratio/low_mean": 0.0006961777307878947, "clip_ratio/low_min": 0.00016920569299448592, "clip_ratio/region_mean": 0.0017766713188999952, "completions/clipped_ratio": 0.0, "completions/max_length": 540.9, "completions/max_terminated_length": 540.9, "completions/mean_length": 201.25443115234376, "completions/mean_terminated_length": 201.25443115234376, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.0288813346841683, "epoch": 24.510560146923783, "frac_reward_zero_std": 0.1500000048428774, "grad_norm": 0.0042724609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4098444818.0, "reward": 7.777604055404663, "reward_std": 0.27822424173355104, "rewards/combined_reward/mean": 7.777603721618652, "rewards/combined_reward/std": 2.5175444364547728, "step": 40050 }, { "clip_ratio/high_max": 0.0019062773698048355, "clip_ratio/high_mean": 0.0011101836100957977, "clip_ratio/low_mean": 0.0006676660049076115, "clip_ratio/low_min": 0.0001681489101126014, "clip_ratio/region_mean": 0.001777849624825952, "completions/clipped_ratio": 0.0, "completions/max_length": 587.7, "completions/max_terminated_length": 587.7, "completions/mean_length": 205.2541702270508, "completions/mean_terminated_length": 205.2541702270508, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.028814984144022067, "epoch": 24.52892561983471, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.00604248046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4101533154.0, "reward": 7.813458442687988, "reward_std": 0.30484359562397, "rewards/combined_reward/mean": 7.813458251953125, "rewards/combined_reward/std": 2.525343132019043, "step": 40080 }, { "clip_ratio/high_max": 0.002012549029313959, "clip_ratio/high_mean": 0.0011617165199519755, "clip_ratio/low_mean": 0.0008050205092634617, "clip_ratio/low_min": 0.0002245419659933153, "clip_ratio/region_mean": 0.0019667370312769585, "completions/clipped_ratio": 0.0, "completions/max_length": 642.7, "completions/max_terminated_length": 642.7, "completions/mean_length": 202.32526702880858, "completions/mean_terminated_length": 202.32526702880858, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.030096208521475393, "epoch": 24.54729109274564, "frac_reward_zero_std": 0.19583334177732467, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4104610243.0, "reward": 8.076270294189452, "reward_std": 0.3532468110322952, "rewards/combined_reward/mean": 8.076269960403442, "rewards/combined_reward/std": 2.2180294156074525, "step": 40110 }, { "clip_ratio/high_max": 0.002116497610889686, "clip_ratio/high_mean": 0.0010923695236366862, "clip_ratio/low_mean": 0.0006863770013296744, "clip_ratio/low_min": 0.00018019506424025167, "clip_ratio/region_mean": 0.00177874652678535, "completions/clipped_ratio": 0.0, "completions/max_length": 585.4, "completions/max_terminated_length": 585.4, "completions/mean_length": 195.64505767822266, "completions/mean_terminated_length": 195.64505767822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.027829985739663242, "epoch": 24.565656565656564, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4107661680.0, "reward": 7.978918123245239, "reward_std": 0.35787140280008317, "rewards/combined_reward/mean": 7.978917646408081, "rewards/combined_reward/std": 2.253950047492981, "step": 40140 }, { "clip_ratio/high_max": 0.0021542290914415693, "clip_ratio/high_mean": 0.0011818316755428289, "clip_ratio/low_mean": 0.0008912683607680568, "clip_ratio/low_min": 0.00019027859622534985, "clip_ratio/region_mean": 0.0020731000215164384, "completions/clipped_ratio": 0.0, "completions/max_length": 564.2, "completions/max_terminated_length": 564.2, "completions/mean_length": 201.26380462646483, "completions/mean_terminated_length": 201.26380462646483, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.030998354032635688, "epoch": 24.584022038567493, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4110734693.0, "reward": 8.012879943847656, "reward_std": 0.28273792266845704, "rewards/combined_reward/mean": 8.012879610061646, "rewards/combined_reward/std": 2.2483651280403136, "step": 40170 }, { "clip_ratio/high_max": 0.0016851146686046074, "clip_ratio/high_mean": 0.0010112467886453185, "clip_ratio/low_mean": 0.0007540572125435574, "clip_ratio/low_min": 0.00020908620936097578, "clip_ratio/region_mean": 0.0017653039933065884, "completions/clipped_ratio": 0.0, "completions/max_length": 567.2, "completions/max_terminated_length": 567.2, "completions/mean_length": 185.36172180175782, "completions/mean_terminated_length": 185.36172180175782, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028238689402739207, "epoch": 24.60238751147842, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4113746642.0, "reward": 8.25643081665039, "reward_std": 0.294978441298008, "rewards/combined_reward/mean": 8.256430387496948, "rewards/combined_reward/std": 2.073008382320404, "step": 40200 }, { "clip_ratio/high_max": 0.0020785041927107767, "clip_ratio/high_mean": 0.001190823327609299, "clip_ratio/low_mean": 0.0007807238160845979, "clip_ratio/low_min": 0.00019801932357950137, "clip_ratio/region_mean": 0.001971547124170077, "completions/clipped_ratio": 0.0, "completions/max_length": 525.9, "completions/max_terminated_length": 525.9, "completions/mean_length": 201.5054733276367, "completions/mean_terminated_length": 201.5054733276367, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.029595204178864756, "epoch": 24.62075298438935, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.00537109375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4116820583.0, "reward": 8.131119775772095, "reward_std": 0.3051419615745544, "rewards/combined_reward/mean": 8.131119585037231, "rewards/combined_reward/std": 2.309945857524872, "step": 40230 }, { "clip_ratio/high_max": 0.0018876124309220661, "clip_ratio/high_mean": 0.0011894949158886448, "clip_ratio/low_mean": 0.000868312486151505, "clip_ratio/low_min": 0.00029349596055302146, "clip_ratio/region_mean": 0.0020578073842140534, "completions/clipped_ratio": 0.0, "completions/max_length": 634.2, "completions/max_terminated_length": 634.2, "completions/mean_length": 195.29115142822266, "completions/mean_terminated_length": 195.29115142822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030717113008722662, "epoch": 24.639118457300274, "frac_reward_zero_std": 0.1708333406597376, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4119870661.0, "reward": 7.949253082275391, "reward_std": 0.3300914645195007, "rewards/combined_reward/mean": 7.949253034591675, "rewards/combined_reward/std": 2.3401187896728515, "step": 40260 }, { "clip_ratio/high_max": 0.0018203290693539506, "clip_ratio/high_mean": 0.001113851949291226, "clip_ratio/low_mean": 0.0007775157548167045, "clip_ratio/low_min": 0.00019649338331267547, "clip_ratio/region_mean": 0.0018913677066545156, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 199.02057952880858, "completions/mean_terminated_length": 199.02057952880858, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02714578352558116, "epoch": 24.657483930211203, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4122935060.0, "reward": 8.010590028762817, "reward_std": 0.33583598881959914, "rewards/combined_reward/mean": 8.010589742660523, "rewards/combined_reward/std": 2.368097484111786, "step": 40290 }, { "clip_ratio/high_max": 0.0021183673471872073, "clip_ratio/high_mean": 0.0011090230585978134, "clip_ratio/low_mean": 0.0007444880992503992, "clip_ratio/low_min": 0.0002369829210996007, "clip_ratio/region_mean": 0.00185351115311884, "completions/clipped_ratio": 0.0, "completions/max_length": 521.1, "completions/max_terminated_length": 521.1, "completions/mean_length": 200.52734832763673, "completions/mean_terminated_length": 200.52734832763673, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028535292902961373, "epoch": 24.67584940312213, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.01043701171875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4126005245.0, "reward": 7.789345169067383, "reward_std": 0.34362357556819917, "rewards/combined_reward/mean": 7.78934497833252, "rewards/combined_reward/std": 2.3559767246246337, "step": 40320 }, { "clip_ratio/high_max": 0.00187223064325129, "clip_ratio/high_mean": 0.0010192326612013859, "clip_ratio/low_mean": 0.0008119832618831424, "clip_ratio/low_min": 0.00020056816016828332, "clip_ratio/region_mean": 0.0018312159193252834, "completions/clipped_ratio": 0.0, "completions/max_length": 590.7, "completions/max_terminated_length": 590.7, "completions/mean_length": 179.77865142822264, "completions/mean_terminated_length": 179.77865142822264, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.028337120497599243, "epoch": 24.694214876033058, "frac_reward_zero_std": 0.16666667424142362, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4128995755.0, "reward": 7.726698207855224, "reward_std": 0.3494249112904072, "rewards/combined_reward/mean": 7.726698112487793, "rewards/combined_reward/std": 2.4224985361099245, "step": 40350 }, { "clip_ratio/high_max": 0.0015999167953850701, "clip_ratio/high_mean": 0.0009444208667749384, "clip_ratio/low_mean": 0.0007451337364424641, "clip_ratio/low_min": 0.00022633225768610524, "clip_ratio/region_mean": 0.0016895545809044658, "completions/clipped_ratio": 0.0, "completions/max_length": 536.2, "completions/max_terminated_length": 536.2, "completions/mean_length": 205.57422332763673, "completions/mean_terminated_length": 205.57422332763673, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.026715828431770207, "epoch": 24.712580348943984, "frac_reward_zero_std": 0.17916666828095912, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4132085320.0, "reward": 7.756226015090943, "reward_std": 0.38620905876159667, "rewards/combined_reward/mean": 7.756225728988648, "rewards/combined_reward/std": 2.60556275844574, "step": 40380 }, { "clip_ratio/high_max": 0.0019657583213605295, "clip_ratio/high_mean": 0.0010308592489309376, "clip_ratio/low_mean": 0.0007766172223152049, "clip_ratio/low_min": 0.00024131394069020946, "clip_ratio/region_mean": 0.0018074764739139937, "completions/clipped_ratio": 0.0, "completions/max_length": 530.1, "completions/max_terminated_length": 530.1, "completions/mean_length": 194.5572967529297, "completions/mean_terminated_length": 194.5572967529297, "completions/min_length": 47.8, "completions/min_terminated_length": 47.8, "entropy": 0.028431346298505862, "epoch": 24.730945821854913, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4135132580.0, "reward": 8.07594838142395, "reward_std": 0.27064715176820753, "rewards/combined_reward/mean": 8.075948238372803, "rewards/combined_reward/std": 2.3405508160591126, "step": 40410 }, { "clip_ratio/high_max": 0.002100445632337748, "clip_ratio/high_mean": 0.0011962195613402097, "clip_ratio/low_mean": 0.0007292538250718887, "clip_ratio/low_min": 0.00020292714398237877, "clip_ratio/region_mean": 0.0019254733818039919, "completions/clipped_ratio": 0.0, "completions/max_length": 628.6, "completions/max_terminated_length": 628.6, "completions/mean_length": 210.57162017822264, "completions/mean_terminated_length": 210.57162017822264, "completions/min_length": 51.8, "completions/min_terminated_length": 51.8, "entropy": 0.030152074821914236, "epoch": 24.74931129476584, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4138241335.0, "reward": 8.177286338806152, "reward_std": 0.3692284166812897, "rewards/combined_reward/mean": 8.177286148071289, "rewards/combined_reward/std": 2.2441174268722532, "step": 40440 }, { "clip_ratio/high_max": 0.0020339815760962664, "clip_ratio/high_mean": 0.0012169572097870212, "clip_ratio/low_mean": 0.0007349215008919904, "clip_ratio/low_min": 0.00020941628996903698, "clip_ratio/region_mean": 0.0019518787133468625, "completions/clipped_ratio": 0.0, "completions/max_length": 634.4, "completions/max_terminated_length": 634.4, "completions/mean_length": 198.2877639770508, "completions/mean_terminated_length": 198.2877639770508, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.02754960278980434, "epoch": 24.767676767676768, "frac_reward_zero_std": 0.2125000074505806, "grad_norm": 0.0458984375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4141302920.0, "reward": 8.097913455963134, "reward_std": 0.28820894509553907, "rewards/combined_reward/mean": 8.097913312911988, "rewards/combined_reward/std": 2.2039858758449555, "step": 40470 }, { "clip_ratio/high_max": 0.0015846881513425614, "clip_ratio/high_mean": 0.0008712720714053527, "clip_ratio/low_mean": 0.0007092299712288271, "clip_ratio/low_min": 0.00018611062259878963, "clip_ratio/region_mean": 0.001580502026627073, "completions/clipped_ratio": 0.0, "completions/max_length": 571.2, "completions/max_terminated_length": 571.2, "completions/mean_length": 201.7401123046875, "completions/mean_terminated_length": 201.7401123046875, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026337066067693135, "epoch": 24.786042240587694, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.02734375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4144377762.0, "reward": 8.326948547363282, "reward_std": 0.3116017386317253, "rewards/combined_reward/mean": 8.326948404312134, "rewards/combined_reward/std": 1.9632122278213502, "step": 40500 }, { "clip_ratio/high_max": 0.0020951207843609154, "clip_ratio/high_mean": 0.0012600512334756786, "clip_ratio/low_mean": 0.0008030700436696254, "clip_ratio/low_min": 0.0002592425599383811, "clip_ratio/region_mean": 0.0020631212668376976, "completions/clipped_ratio": 0.0, "completions/max_length": 682.6, "completions/max_terminated_length": 682.6, "completions/mean_length": 205.91927642822264, "completions/mean_terminated_length": 205.91927642822264, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.030328595250224074, "epoch": 24.804407713498623, "frac_reward_zero_std": 0.17083333507180215, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4147468652.0, "reward": 7.877101039886474, "reward_std": 0.2982399001717567, "rewards/combined_reward/mean": 7.877100849151612, "rewards/combined_reward/std": 2.5828229665756224, "step": 40530 }, { "clip_ratio/high_max": 0.0019174072042612048, "clip_ratio/high_mean": 0.001103967362723779, "clip_ratio/low_mean": 0.0007399334292131244, "clip_ratio/low_min": 0.0002232956188284637, "clip_ratio/region_mean": 0.0018439007981214672, "completions/clipped_ratio": 0.0, "completions/max_length": 607.5, "completions/max_terminated_length": 607.5, "completions/mean_length": 206.01068115234375, "completions/mean_terminated_length": 206.01068115234375, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.029202839297552904, "epoch": 24.82277318640955, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4150559893.0, "reward": 7.988489151000977, "reward_std": 0.33177229315042495, "rewards/combined_reward/mean": 7.988488960266113, "rewards/combined_reward/std": 2.087195897102356, "step": 40560 }, { "clip_ratio/high_max": 0.0020117006943716357, "clip_ratio/high_mean": 0.0010998160141753033, "clip_ratio/low_mean": 0.0008217783836395635, "clip_ratio/low_min": 0.00020251433161320165, "clip_ratio/region_mean": 0.0019215944048482925, "completions/clipped_ratio": 0.0, "completions/max_length": 539.1, "completions/max_terminated_length": 539.1, "completions/mean_length": 189.73047332763673, "completions/mean_terminated_length": 189.73047332763673, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029037315755461653, "epoch": 24.841138659320478, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.003387451171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4153588618.0, "reward": 7.9612708568573, "reward_std": 0.29053641110658646, "rewards/combined_reward/mean": 7.961270475387574, "rewards/combined_reward/std": 2.380772602558136, "step": 40590 }, { "clip_ratio/high_max": 0.0015159641392529012, "clip_ratio/high_mean": 0.0008679744649271015, "clip_ratio/low_mean": 0.0006823003641329706, "clip_ratio/low_min": 0.00021908774215262383, "clip_ratio/region_mean": 0.0015502748317279232, "completions/clipped_ratio": 0.0, "completions/max_length": 656.6, "completions/max_terminated_length": 656.6, "completions/mean_length": 202.2369827270508, "completions/mean_terminated_length": 202.2369827270508, "completions/min_length": 47.1, "completions/min_terminated_length": 47.1, "entropy": 0.02480712824811538, "epoch": 24.859504132231404, "frac_reward_zero_std": 0.13750000447034835, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4156665368.0, "reward": 8.023663139343261, "reward_std": 0.322966206073761, "rewards/combined_reward/mean": 8.023662757873534, "rewards/combined_reward/std": 2.43300656080246, "step": 40620 }, { "clip_ratio/high_max": 0.0019663283451033446, "clip_ratio/high_mean": 0.00113319101704595, "clip_ratio/low_mean": 0.0006479250693397868, "clip_ratio/low_min": 0.00018769964735838586, "clip_ratio/region_mean": 0.0017811160990580295, "completions/clipped_ratio": 0.0, "completions/max_length": 602.5, "completions/max_terminated_length": 602.5, "completions/mean_length": 198.3000015258789, "completions/mean_terminated_length": 198.3000015258789, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.0315725412995865, "epoch": 24.877869605142333, "frac_reward_zero_std": 0.20833334177732468, "grad_norm": 0.024169921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4159727000.0, "reward": 8.09168095588684, "reward_std": 0.34221314191818236, "rewards/combined_reward/mean": 8.091680908203125, "rewards/combined_reward/std": 2.2573223233222963, "step": 40650 }, { "clip_ratio/high_max": 0.0018554245849372818, "clip_ratio/high_mean": 0.0010441260431737949, "clip_ratio/low_mean": 0.0009041694547098207, "clip_ratio/low_min": 0.0003136046613993434, "clip_ratio/region_mean": 0.0019482955108590734, "completions/clipped_ratio": 0.0, "completions/max_length": 583.3, "completions/max_terminated_length": 583.3, "completions/mean_length": 205.4166732788086, "completions/mean_terminated_length": 205.4166732788086, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.02979116518981755, "epoch": 24.89623507805326, "frac_reward_zero_std": 0.16666666902601718, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4162815960.0, "reward": 8.136965227127074, "reward_std": 0.28347503393888474, "rewards/combined_reward/mean": 8.13696517944336, "rewards/combined_reward/std": 2.118568253517151, "step": 40680 }, { "clip_ratio/high_max": 0.0018284487698110752, "clip_ratio/high_mean": 0.0010196655823771531, "clip_ratio/low_mean": 0.0008255291992099956, "clip_ratio/low_min": 0.00018911891966126858, "clip_ratio/region_mean": 0.001845194774796255, "completions/clipped_ratio": 0.0, "completions/max_length": 592.1, "completions/max_terminated_length": 592.1, "completions/mean_length": 205.27474365234374, "completions/mean_terminated_length": 205.27474365234374, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.0298396077627937, "epoch": 24.914600550964188, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.006072998046875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4165904375.0, "reward": 7.910553407669068, "reward_std": 0.3628436326980591, "rewards/combined_reward/mean": 7.910553312301635, "rewards/combined_reward/std": 2.2388519287109374, "step": 40710 }, { "clip_ratio/high_max": 0.0019420797907514498, "clip_ratio/high_mean": 0.001076427849814839, "clip_ratio/low_mean": 0.0007134970600115291, "clip_ratio/low_min": 0.00021002601788495668, "clip_ratio/region_mean": 0.0017899249083711765, "completions/clipped_ratio": 0.0, "completions/max_length": 553.7, "completions/max_terminated_length": 553.7, "completions/mean_length": 197.07917022705078, "completions/mean_terminated_length": 197.07917022705078, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.027912684095402557, "epoch": 24.932966023875114, "frac_reward_zero_std": 0.1750000063329935, "grad_norm": 0.00689697265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4168961319.0, "reward": 8.03018307685852, "reward_std": 0.28743835389614103, "rewards/combined_reward/mean": 8.030182647705079, "rewards/combined_reward/std": 2.246380400657654, "step": 40740 }, { "clip_ratio/high_max": 0.0017631054268955873, "clip_ratio/high_mean": 0.0010463713725281801, "clip_ratio/low_mean": 0.0006690810504248172, "clip_ratio/low_min": 0.0001759635114770693, "clip_ratio/region_mean": 0.0017154524110083002, "completions/clipped_ratio": 0.0, "completions/max_length": 601.4, "completions/max_terminated_length": 601.4, "completions/mean_length": 210.27344360351563, "completions/mean_terminated_length": 210.27344360351563, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.026908402719224494, "epoch": 24.951331496786043, "frac_reward_zero_std": 0.16250000707805157, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4172068929.0, "reward": 8.07895998954773, "reward_std": 0.32581260800361633, "rewards/combined_reward/mean": 8.078959655761718, "rewards/combined_reward/std": 2.077314567565918, "step": 40770 }, { "clip_ratio/high_max": 0.0018288227865317216, "clip_ratio/high_mean": 0.001058591465941087, "clip_ratio/low_mean": 0.0007315640469945113, "clip_ratio/low_min": 0.00016864284722638938, "clip_ratio/region_mean": 0.0017901555072360983, "completions/clipped_ratio": 0.0, "completions/max_length": 538.6, "completions/max_terminated_length": 538.6, "completions/mean_length": 207.49818267822266, "completions/mean_terminated_length": 207.49818267822266, "completions/min_length": 50.8, "completions/min_terminated_length": 50.8, "entropy": 0.02915126715476314, "epoch": 24.96969696969697, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.006622314453125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4175165882.0, "reward": 8.060931491851807, "reward_std": 0.2965683326125145, "rewards/combined_reward/mean": 8.060931396484374, "rewards/combined_reward/std": 2.166172993183136, "step": 40800 }, { "clip_ratio/high_max": 0.0019527702813017338, "clip_ratio/high_mean": 0.0011355276910762768, "clip_ratio/low_mean": 0.0008723914594156668, "clip_ratio/low_min": 0.0002822636580579759, "clip_ratio/region_mean": 0.002007919150249412, "completions/clipped_ratio": 0.0, "completions/max_length": 617.7, "completions/max_terminated_length": 617.7, "completions/mean_length": 196.4078170776367, "completions/mean_terminated_length": 196.4078170776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.0285131783845524, "epoch": 24.988062442607898, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4178220248.0, "reward": 8.146965599060058, "reward_std": 0.32688435167074203, "rewards/combined_reward/mean": 8.14696536064148, "rewards/combined_reward/std": 2.173694336414337, "step": 40830 }, { "clip_ratio/high_max": 0.0019400843288405935, "clip_ratio/high_mean": 0.0011049766195713748, "clip_ratio/low_mean": 0.0006907274067491583, "clip_ratio/low_min": 0.00018351394268035216, "clip_ratio/region_mean": 0.0017957040390236395, "completions/clipped_ratio": 0.0, "completions/max_length": 624.4, "completions/max_terminated_length": 624.4, "completions/mean_length": 207.8713577270508, "completions/mean_terminated_length": 207.8713577270508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.026598813050411275, "epoch": 25.006734006734007, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4181318634.0, "reward": 8.068601417541505, "reward_std": 0.3390425696969032, "rewards/combined_reward/mean": 8.068601417541505, "rewards/combined_reward/std": 2.100806248188019, "step": 40860 }, { "clip_ratio/high_max": 0.002323574279337966, "clip_ratio/high_mean": 0.0013277713495578305, "clip_ratio/low_mean": 0.0009307710680028929, "clip_ratio/low_min": 0.00022609446338416698, "clip_ratio/region_mean": 0.0022585424240484523, "completions/clipped_ratio": 0.0, "completions/max_length": 581.4, "completions/max_terminated_length": 581.4, "completions/mean_length": 201.65677642822266, "completions/mean_terminated_length": 201.65677642822266, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02629634163652857, "epoch": 25.025099479644933, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4184393156.0, "reward": 8.307684421539307, "reward_std": 0.2860836789011955, "rewards/combined_reward/mean": 8.307683801651, "rewards/combined_reward/std": 2.0566641211509706, "step": 40890 }, { "clip_ratio/high_max": 0.00231285010886495, "clip_ratio/high_mean": 0.0013231076843415698, "clip_ratio/low_mean": 0.0008117579277798844, "clip_ratio/low_min": 0.00017805782626965084, "clip_ratio/region_mean": 0.0021348656203675396, "completions/clipped_ratio": 0.0, "completions/max_length": 588.5, "completions/max_terminated_length": 588.5, "completions/mean_length": 201.55964202880858, "completions/mean_terminated_length": 201.55964202880858, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.026257605059072375, "epoch": 25.043464952555862, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4187467305.0, "reward": 8.015105199813842, "reward_std": 0.3825759254395962, "rewards/combined_reward/mean": 8.01510500907898, "rewards/combined_reward/std": 2.273582029342651, "step": 40920 }, { "clip_ratio/high_max": 0.0028130724963072377, "clip_ratio/high_mean": 0.0014469206626624024, "clip_ratio/low_mean": 0.0008693478398830242, "clip_ratio/low_min": 0.0002067802861953775, "clip_ratio/region_mean": 0.002316268497573522, "completions/clipped_ratio": 0.0, "completions/max_length": 539.3, "completions/max_terminated_length": 539.3, "completions/mean_length": 197.07709045410155, "completions/mean_terminated_length": 197.07709045410155, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.0281272758419315, "epoch": 25.061830425466788, "frac_reward_zero_std": 0.17500000186264514, "grad_norm": 0.005706787109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4190524241.0, "reward": 8.026017427444458, "reward_std": 0.3660419017076492, "rewards/combined_reward/mean": 8.026017093658448, "rewards/combined_reward/std": 2.310856056213379, "step": 40950 }, { "clip_ratio/high_max": 0.002141612986694478, "clip_ratio/high_mean": 0.0011902282451046631, "clip_ratio/low_mean": 0.0008013971693799249, "clip_ratio/low_min": 0.00017907529351456712, "clip_ratio/region_mean": 0.0019916254226700404, "completions/clipped_ratio": 0.0, "completions/max_length": 599.5, "completions/max_terminated_length": 599.5, "completions/mean_length": 210.0697982788086, "completions/mean_terminated_length": 210.0697982788086, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02600375396820406, "epoch": 25.080195898377717, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.00762939453125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4193631069.0, "reward": 8.168495416641235, "reward_std": 0.3472735434770584, "rewards/combined_reward/mean": 8.168495082855225, "rewards/combined_reward/std": 2.12459682226181, "step": 40980 }, { "clip_ratio/high_max": 0.002366614419831118, "clip_ratio/high_mean": 0.0013845431309164268, "clip_ratio/low_mean": 0.0008179135801280305, "clip_ratio/low_min": 0.00023938405647641047, "clip_ratio/region_mean": 0.0022024567316596706, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 203.61407012939452, "completions/mean_terminated_length": 203.61407012939452, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.026172424387186765, "epoch": 25.098561371288643, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.00665283203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4196713107.0, "reward": 8.11976432800293, "reward_std": 0.314256364107132, "rewards/combined_reward/mean": 8.119763898849488, "rewards/combined_reward/std": 2.4408764123916624, "step": 41010 }, { "clip_ratio/high_max": 0.0024211131041132223, "clip_ratio/high_mean": 0.0013681314458760122, "clip_ratio/low_mean": 0.0010784030164359137, "clip_ratio/low_min": 0.0003195224031514954, "clip_ratio/region_mean": 0.0024465344739534583, "completions/clipped_ratio": 0.0, "completions/max_length": 652.9, "completions/max_terminated_length": 652.9, "completions/mean_length": 205.54349517822266, "completions/mean_terminated_length": 205.54349517822266, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03262298411379258, "epoch": 25.116926844199572, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4199802554.0, "reward": 8.219353437423706, "reward_std": 0.3246060013771057, "rewards/combined_reward/mean": 8.21935315132141, "rewards/combined_reward/std": 1.9557159543037415, "step": 41040 }, { "clip_ratio/high_max": 0.0022834992356365546, "clip_ratio/high_mean": 0.001358085570973344, "clip_ratio/low_mean": 0.0010198017958221802, "clip_ratio/low_min": 0.0002829997746933562, "clip_ratio/region_mean": 0.0023778873655828647, "completions/clipped_ratio": 0.0, "completions/max_length": 625.6, "completions/max_terminated_length": 625.6, "completions/mean_length": 213.9166702270508, "completions/mean_terminated_length": 213.9166702270508, "completions/min_length": 47.0, "completions/min_terminated_length": 47.0, "entropy": 0.02943820416306456, "epoch": 25.135292317110498, "frac_reward_zero_std": 0.17083333991467953, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4202924154.0, "reward": 8.129235649108887, "reward_std": 0.30482603162527083, "rewards/combined_reward/mean": 8.129235458374023, "rewards/combined_reward/std": 2.232584524154663, "step": 41070 }, { "clip_ratio/high_max": 0.00234050352689034, "clip_ratio/high_mean": 0.0013528206614864756, "clip_ratio/low_mean": 0.0008838373138132738, "clip_ratio/low_min": 0.0002274171255218486, "clip_ratio/region_mean": 0.00223665797408709, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 203.12240142822264, "completions/mean_terminated_length": 203.12240142822264, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.02810657360435774, "epoch": 25.153657790021427, "frac_reward_zero_std": 0.18333333656191825, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4206004304.0, "reward": 8.14833960533142, "reward_std": 0.2915416195988655, "rewards/combined_reward/mean": 8.148339414596558, "rewards/combined_reward/std": 2.0983516693115236, "step": 41100 }, { "clip_ratio/high_max": 0.0019185677857118813, "clip_ratio/high_mean": 0.001075244071034831, "clip_ratio/low_mean": 0.0008357236574132305, "clip_ratio/low_min": 0.0002217571178334765, "clip_ratio/region_mean": 0.0019109677436063067, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 201.7018280029297, "completions/mean_terminated_length": 201.7018280029297, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02478405909302334, "epoch": 25.172023262932353, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.00616455078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4209078999.0, "reward": 8.228160858154297, "reward_std": 0.3209026038646698, "rewards/combined_reward/mean": 8.228160762786866, "rewards/combined_reward/std": 2.0603532195091248, "step": 41130 }, { "clip_ratio/high_max": 0.0025941949512343853, "clip_ratio/high_mean": 0.0014318002467916813, "clip_ratio/low_mean": 0.0010943697287681667, "clip_ratio/low_min": 0.0002747486893592092, "clip_ratio/region_mean": 0.0025261699687689544, "completions/clipped_ratio": 0.0, "completions/max_length": 637.3, "completions/max_terminated_length": 637.3, "completions/mean_length": 202.33047485351562, "completions/mean_terminated_length": 202.33047485351562, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03079002907810112, "epoch": 25.190388735843282, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4212156108.0, "reward": 8.044478034973144, "reward_std": 0.30362043529748917, "rewards/combined_reward/mean": 8.044477701187134, "rewards/combined_reward/std": 2.4437608122825623, "step": 41160 }, { "clip_ratio/high_max": 0.0024452888891876986, "clip_ratio/high_mean": 0.0014455327793257311, "clip_ratio/low_mean": 0.0008284358477491575, "clip_ratio/low_min": 0.00026052698764639595, "clip_ratio/region_mean": 0.002273968634350846, "completions/clipped_ratio": 0.0, "completions/max_length": 569.6, "completions/max_terminated_length": 569.6, "completions/mean_length": 193.96146240234376, "completions/mean_terminated_length": 193.96146240234376, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.02709313854575157, "epoch": 25.208754208754208, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4215201080.0, "reward": 8.29511342048645, "reward_std": 0.2995273917913437, "rewards/combined_reward/mean": 8.295113039016723, "rewards/combined_reward/std": 2.017604315280914, "step": 41190 }, { "clip_ratio/high_max": 0.0022030230296271234, "clip_ratio/high_mean": 0.0011931193021155195, "clip_ratio/low_mean": 0.0008308588077246289, "clip_ratio/low_min": 0.00020103504890964057, "clip_ratio/region_mean": 0.002023978105474574, "completions/clipped_ratio": 0.0, "completions/max_length": 648.9, "completions/max_terminated_length": 648.9, "completions/mean_length": 196.58021087646483, "completions/mean_terminated_length": 196.58021087646483, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.025024638907052577, "epoch": 25.227119681665137, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.006317138671875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4218256108.0, "reward": 7.918447256088257, "reward_std": 0.3698273330926895, "rewards/combined_reward/mean": 7.918447208404541, "rewards/combined_reward/std": 2.405626058578491, "step": 41220 }, { "clip_ratio/high_max": 0.002114480228434938, "clip_ratio/high_mean": 0.001260452593366305, "clip_ratio/low_mean": 0.0008653002806871276, "clip_ratio/low_min": 0.00015671555617397342, "clip_ratio/region_mean": 0.0021257528860587627, "completions/clipped_ratio": 0.0, "completions/max_length": 539.7, "completions/max_terminated_length": 539.7, "completions/mean_length": 195.63906707763672, "completions/mean_terminated_length": 195.63906707763672, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.02554652062244713, "epoch": 25.245485154576063, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4221307522.0, "reward": 8.043769311904907, "reward_std": 0.35578196346759794, "rewards/combined_reward/mean": 8.043768978118896, "rewards/combined_reward/std": 2.198091518878937, "step": 41250 }, { "clip_ratio/high_max": 0.0028450541838537903, "clip_ratio/high_mean": 0.001556106826562124, "clip_ratio/low_mean": 0.0008711983978476685, "clip_ratio/low_min": 0.00013932706060586496, "clip_ratio/region_mean": 0.0024273052064624303, "completions/clipped_ratio": 0.0, "completions/max_length": 584.5, "completions/max_terminated_length": 584.5, "completions/mean_length": 198.5213607788086, "completions/mean_terminated_length": 198.5213607788086, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.03423594476965566, "epoch": 25.263850627486992, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.033935546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4224370004.0, "reward": 8.1663902759552, "reward_std": 0.3539436191320419, "rewards/combined_reward/mean": 8.166389894485473, "rewards/combined_reward/std": 2.1258671402931215, "step": 41280 }, { "clip_ratio/high_max": 0.0026975369272501364, "clip_ratio/high_mean": 0.0014744398951734184, "clip_ratio/low_mean": 0.0008382974924946514, "clip_ratio/low_min": 0.00023564421280752867, "clip_ratio/region_mean": 0.0023127373851214847, "completions/clipped_ratio": 0.0, "completions/max_length": 550.3, "completions/max_terminated_length": 550.3, "completions/mean_length": 197.85443420410155, "completions/mean_terminated_length": 197.85443420410155, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.027989238686859607, "epoch": 25.282216100397918, "frac_reward_zero_std": 0.2041666716337204, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4227429925.0, "reward": 7.992238140106201, "reward_std": 0.300083139538765, "rewards/combined_reward/mean": 7.992237901687622, "rewards/combined_reward/std": 2.266592872142792, "step": 41310 }, { "clip_ratio/high_max": 0.002517390629994528, "clip_ratio/high_mean": 0.0013967892195069, "clip_ratio/low_mean": 0.0008119812371054044, "clip_ratio/low_min": 0.00018543306384041596, "clip_ratio/region_mean": 0.002208770470072826, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 200.09271545410155, "completions/mean_terminated_length": 200.09271545410155, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.027272005433527133, "epoch": 25.300581573308847, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4230498441.0, "reward": 8.094974565505982, "reward_std": 0.3338675931096077, "rewards/combined_reward/mean": 8.094974327087403, "rewards/combined_reward/std": 2.264859712123871, "step": 41340 }, { "clip_ratio/high_max": 0.0031761235562347188, "clip_ratio/high_mean": 0.0017327723360722302, "clip_ratio/low_mean": 0.0008372782659231841, "clip_ratio/low_min": 0.0001797107353922911, "clip_ratio/region_mean": 0.0025700505919909727, "completions/clipped_ratio": 0.0, "completions/max_length": 687.8, "completions/max_terminated_length": 687.8, "completions/mean_length": 192.81614990234374, "completions/mean_terminated_length": 192.81614990234374, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.036072738468647, "epoch": 25.318947046219773, "frac_reward_zero_std": 0.20416667126119137, "grad_norm": 0.007659912109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4233539015.0, "reward": 7.999736499786377, "reward_std": 0.3129606157541275, "rewards/combined_reward/mean": 7.999736309051514, "rewards/combined_reward/std": 2.4041842103004454, "step": 41370 }, { "clip_ratio/high_max": 0.002200637842179276, "clip_ratio/high_mean": 0.0013373293988479418, "clip_ratio/low_mean": 0.0010645299372602798, "clip_ratio/low_min": 0.0003159767688581875, "clip_ratio/region_mean": 0.0024018593234359288, "completions/clipped_ratio": 0.0, "completions/max_length": 548.6, "completions/max_terminated_length": 548.6, "completions/mean_length": 198.2190155029297, "completions/mean_terminated_length": 198.2190155029297, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028618016311277947, "epoch": 25.337312519130702, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.014892578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4236600336.0, "reward": 8.249741411209106, "reward_std": 0.25765300914645195, "rewards/combined_reward/mean": 8.249741458892823, "rewards/combined_reward/std": 2.030086004734039, "step": 41400 }, { "clip_ratio/high_max": 0.0026713188577559776, "clip_ratio/high_mean": 0.00155060008255532, "clip_ratio/low_mean": 0.0009124340068713839, "clip_ratio/low_min": 0.0001824844109554154, "clip_ratio/region_mean": 0.00246303409464114, "completions/clipped_ratio": 0.0, "completions/max_length": 607.9, "completions/max_terminated_length": 607.9, "completions/mean_length": 204.77865142822264, "completions/mean_terminated_length": 204.77865142822264, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.030236867846300204, "epoch": 25.355677992041628, "frac_reward_zero_std": 0.18333333507180213, "grad_norm": 0.00634765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4239686846.0, "reward": 8.192913103103638, "reward_std": 0.3394464686512947, "rewards/combined_reward/mean": 8.192912817001343, "rewards/combined_reward/std": 2.1476980686187743, "step": 41430 }, { "clip_ratio/high_max": 0.0022434404391484955, "clip_ratio/high_mean": 0.0012784092031021524, "clip_ratio/low_mean": 0.0009333005452087187, "clip_ratio/low_min": 0.00023409728722375198, "clip_ratio/region_mean": 0.002211709763893547, "completions/clipped_ratio": 0.0, "completions/max_length": 600.4, "completions/max_terminated_length": 600.4, "completions/mean_length": 195.31328582763672, "completions/mean_terminated_length": 195.31328582763672, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.029206417811413605, "epoch": 25.374043464952557, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0240478515625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4242737009.0, "reward": 8.2345627784729, "reward_std": 0.34027910977602005, "rewards/combined_reward/mean": 8.234562540054322, "rewards/combined_reward/std": 2.086388885974884, "step": 41460 }, { "clip_ratio/high_max": 0.00217379549042865, "clip_ratio/high_mean": 0.001276698981746449, "clip_ratio/low_mean": 0.0008846821923119327, "clip_ratio/low_min": 0.00020548822406757, "clip_ratio/region_mean": 0.0021613811783026903, "completions/clipped_ratio": 0.0, "completions/max_length": 485.5, "completions/max_terminated_length": 485.5, "completions/mean_length": 189.84870300292968, "completions/mean_terminated_length": 189.84870300292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02601363992628952, "epoch": 25.392408937863483, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 4245766188.0, "reward": 8.30578293800354, "reward_std": 0.2331622928380966, "rewards/combined_reward/mean": 8.30578293800354, "rewards/combined_reward/std": 2.025904381275177, "step": 41490 }, { "clip_ratio/high_max": 0.0025059330117073843, "clip_ratio/high_mean": 0.0013590710185477897, "clip_ratio/low_mean": 0.0008971211675088853, "clip_ratio/low_min": 0.00025762398702984984, "clip_ratio/region_mean": 0.002256192185935409, "completions/clipped_ratio": 0.0, "completions/max_length": 635.1, "completions/max_terminated_length": 635.1, "completions/mean_length": 189.7619827270508, "completions/mean_terminated_length": 189.7619827270508, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.027413502149283887, "epoch": 25.410774410774412, "frac_reward_zero_std": 0.1833333417773247, "grad_norm": 0.03515625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4248795034.0, "reward": 8.195920944213867, "reward_std": 0.3172381728887558, "rewards/combined_reward/mean": 8.195920658111572, "rewards/combined_reward/std": 1.9794097304344178, "step": 41520 }, { "clip_ratio/high_max": 0.002557108527495681, "clip_ratio/high_mean": 0.0014072719752827349, "clip_ratio/low_mean": 0.0009766530252818483, "clip_ratio/low_min": 0.0002020825750757164, "clip_ratio/region_mean": 0.0023839250270005627, "completions/clipped_ratio": 0.0, "completions/max_length": 541.7, "completions/max_terminated_length": 541.7, "completions/mean_length": 199.94688110351564, "completions/mean_terminated_length": 199.94688110351564, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02828369919055452, "epoch": 25.429139883685338, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.006805419921875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4251862990.0, "reward": 8.006407260894775, "reward_std": 0.26571316868066785, "rewards/combined_reward/mean": 8.006407022476196, "rewards/combined_reward/std": 2.482809317111969, "step": 41550 }, { "clip_ratio/high_max": 0.0024515268038764284, "clip_ratio/high_mean": 0.0013511120963812574, "clip_ratio/low_mean": 0.0008999011247700158, "clip_ratio/low_min": 0.00020497848405890787, "clip_ratio/region_mean": 0.002251013240311295, "completions/clipped_ratio": 0.0, "completions/max_length": 537.4, "completions/max_terminated_length": 537.4, "completions/mean_length": 194.32865142822266, "completions/mean_terminated_length": 194.32865142822266, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.027765258005820213, "epoch": 25.447505356596267, "frac_reward_zero_std": 0.18750000633299352, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4254909372.0, "reward": 7.928602933883667, "reward_std": 0.31214811056852343, "rewards/combined_reward/mean": 7.928602695465088, "rewards/combined_reward/std": 2.4210470795631407, "step": 41580 }, { "clip_ratio/high_max": 0.0024623912215853733, "clip_ratio/high_mean": 0.0013695993129658746, "clip_ratio/low_mean": 0.000904500934969595, "clip_ratio/low_min": 0.0002635029658752804, "clip_ratio/region_mean": 0.002274100224894937, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 197.4703170776367, "completions/mean_terminated_length": 197.4703170776367, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.02793920070398599, "epoch": 25.465870829507192, "frac_reward_zero_std": 0.14583333618938923, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4257967818.0, "reward": 7.996175146102905, "reward_std": 0.2741954490542412, "rewards/combined_reward/mean": 7.996174716949463, "rewards/combined_reward/std": 2.3819100975990297, "step": 41610 }, { "clip_ratio/high_max": 0.00288527962960264, "clip_ratio/high_mean": 0.0016013513806683477, "clip_ratio/low_mean": 0.001014459676844126, "clip_ratio/low_min": 0.00022760614495685635, "clip_ratio/region_mean": 0.0026158110612110857, "completions/clipped_ratio": 0.0, "completions/max_length": 605.7, "completions/max_terminated_length": 605.7, "completions/mean_length": 198.95208892822265, "completions/mean_terminated_length": 198.95208892822265, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03036658513980607, "epoch": 25.48423630241812, "frac_reward_zero_std": 0.1833333358168602, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4261031954.0, "reward": 8.310655450820922, "reward_std": 0.340054889023304, "rewards/combined_reward/mean": 8.310655069351196, "rewards/combined_reward/std": 2.128911519050598, "step": 41640 }, { "clip_ratio/high_max": 0.0021609843332650296, "clip_ratio/high_mean": 0.0011150769804468533, "clip_ratio/low_mean": 0.0007804732093063648, "clip_ratio/low_min": 0.00023913587647257374, "clip_ratio/region_mean": 0.0018955501868428352, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 189.98542022705078, "completions/mean_terminated_length": 189.98542022705078, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.024605679574112096, "epoch": 25.502601775329047, "frac_reward_zero_std": 0.18333333693444728, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4264061658.0, "reward": 7.896659088134766, "reward_std": 0.34509294480085373, "rewards/combined_reward/mean": 7.896658945083618, "rewards/combined_reward/std": 2.4831191182136534, "step": 41670 }, { "clip_ratio/high_max": 0.0025666682258209525, "clip_ratio/high_mean": 0.0014853343105035795, "clip_ratio/low_mean": 0.001087695914369154, "clip_ratio/low_min": 0.0002988572299121491, "clip_ratio/region_mean": 0.0025730302072285364, "completions/clipped_ratio": 0.0, "completions/max_length": 625.1, "completions/max_terminated_length": 625.1, "completions/mean_length": 197.6307357788086, "completions/mean_terminated_length": 197.6307357788086, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03491972843185067, "epoch": 25.520967248239977, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4267120720.0, "reward": 7.934893465042114, "reward_std": 0.23978697508573532, "rewards/combined_reward/mean": 7.934893178939819, "rewards/combined_reward/std": 2.5224226474761964, "step": 41700 }, { "clip_ratio/high_max": 0.0021296727146060826, "clip_ratio/high_mean": 0.001147582625223246, "clip_ratio/low_mean": 0.0007780005017896959, "clip_ratio/low_min": 0.00017955476481195848, "clip_ratio/region_mean": 0.001925583114886346, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 196.33021392822266, "completions/mean_terminated_length": 196.33021392822266, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.025560414418578148, "epoch": 25.539332721150902, "frac_reward_zero_std": 0.22500000894069672, "grad_norm": 0.00616455078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4270174788.0, "reward": 8.282195901870727, "reward_std": 0.2800623282790184, "rewards/combined_reward/mean": 8.282195520401, "rewards/combined_reward/std": 1.9500994563102723, "step": 41730 }, { "clip_ratio/high_max": 0.0031397716355665278, "clip_ratio/high_mean": 0.0017100281120898822, "clip_ratio/low_mean": 0.0009739063902695004, "clip_ratio/low_min": 0.000289584513908873, "clip_ratio/region_mean": 0.002683934531523846, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 202.31771240234374, "completions/mean_terminated_length": 202.31771240234374, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.030039506669466694, "epoch": 25.55769819406183, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4273251848.0, "reward": 8.134464454650878, "reward_std": 0.3161551907658577, "rewards/combined_reward/mean": 8.134464216232299, "rewards/combined_reward/std": 2.1551761388778687, "step": 41760 }, { "clip_ratio/high_max": 0.0022802210650600804, "clip_ratio/high_mean": 0.0012595742010186465, "clip_ratio/low_mean": 0.0009163568674315077, "clip_ratio/low_min": 0.00020790043272427282, "clip_ratio/region_mean": 0.002175931078090798, "completions/clipped_ratio": 0.0, "completions/max_length": 626.5, "completions/max_terminated_length": 626.5, "completions/mean_length": 202.6515670776367, "completions/mean_terminated_length": 202.6515670776367, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.027485752229889235, "epoch": 25.576063666972757, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4276330190.0, "reward": 8.041602945327758, "reward_std": 0.30577731430530547, "rewards/combined_reward/mean": 8.0416024684906, "rewards/combined_reward/std": 2.0919049859046934, "step": 41790 }, { "clip_ratio/high_max": 0.0021539654500278022, "clip_ratio/high_mean": 0.0012334973293036456, "clip_ratio/low_mean": 0.0009342425937575172, "clip_ratio/low_min": 0.0003018320244639957, "clip_ratio/region_mean": 0.002167739930155221, "completions/clipped_ratio": 0.0, "completions/max_length": 518.6, "completions/max_terminated_length": 518.6, "completions/mean_length": 199.40287170410156, "completions/mean_terminated_length": 199.40287170410156, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.026803566391269365, "epoch": 25.594429139883687, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.0140380859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4279396057.0, "reward": 8.246305084228515, "reward_std": 0.30566649734973905, "rewards/combined_reward/mean": 8.246304941177367, "rewards/combined_reward/std": 2.179079020023346, "step": 41820 }, { "clip_ratio/high_max": 0.0024853917411140477, "clip_ratio/high_mean": 0.0013986404302462083, "clip_ratio/low_mean": 0.0007163662562864677, "clip_ratio/low_min": 0.0002145944941730704, "clip_ratio/region_mean": 0.0021150066970828147, "completions/clipped_ratio": 0.0, "completions/max_length": 584.9, "completions/max_terminated_length": 584.9, "completions/mean_length": 200.15833892822266, "completions/mean_terminated_length": 200.15833892822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026210688621116182, "epoch": 25.612794612794612, "frac_reward_zero_std": 0.1708333395421505, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4282464825.0, "reward": 8.271352863311767, "reward_std": 0.3349998578429222, "rewards/combined_reward/mean": 8.271352624893188, "rewards/combined_reward/std": 1.9777267813682555, "step": 41850 }, { "clip_ratio/high_max": 0.002886220487804773, "clip_ratio/high_mean": 0.0016584984629541093, "clip_ratio/low_mean": 0.0009955335931105462, "clip_ratio/low_min": 0.00027389532454738703, "clip_ratio/region_mean": 0.002654032063340613, "completions/clipped_ratio": 0.0, "completions/max_length": 531.8, "completions/max_terminated_length": 531.8, "completions/mean_length": 193.19427795410155, "completions/mean_terminated_length": 193.19427795410155, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.03100763917900622, "epoch": 25.63116008570554, "frac_reward_zero_std": 0.1791666679084301, "grad_norm": 0.0038604736328125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4285506851.0, "reward": 7.786850643157959, "reward_std": 0.3234362915158272, "rewards/combined_reward/mean": 7.786850500106811, "rewards/combined_reward/std": 2.3997554779052734, "step": 41880 }, { "clip_ratio/high_max": 0.0021533798935706727, "clip_ratio/high_mean": 0.0012017173812637338, "clip_ratio/low_mean": 0.0009713860029781548, "clip_ratio/low_min": 0.0002467785438057035, "clip_ratio/region_mean": 0.0021731033654456646, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 203.37188262939452, "completions/mean_terminated_length": 203.37188262939452, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.025243854887473088, "epoch": 25.649525558616467, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4288587959.0, "reward": 8.007487154006958, "reward_std": 0.2799578353762627, "rewards/combined_reward/mean": 8.007487058639526, "rewards/combined_reward/std": 2.2201871514320373, "step": 41910 }, { "clip_ratio/high_max": 0.002563522373869394, "clip_ratio/high_mean": 0.0013983120843477082, "clip_ratio/low_mean": 0.001088097562994032, "clip_ratio/low_min": 0.0003227441127819475, "clip_ratio/region_mean": 0.0024864096340024844, "completions/clipped_ratio": 0.0, "completions/max_length": 536.2, "completions/max_terminated_length": 536.2, "completions/mean_length": 193.74817962646483, "completions/mean_terminated_length": 193.74817962646483, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.029543117728705208, "epoch": 25.667891031527397, "frac_reward_zero_std": 0.1625000022351742, "grad_norm": 0.0277099609375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4291632112.0, "reward": 8.098860168457032, "reward_std": 0.317243766784668, "rewards/combined_reward/mean": 8.098860120773315, "rewards/combined_reward/std": 2.1898394584655763, "step": 41940 }, { "clip_ratio/high_max": 0.00229346755319663, "clip_ratio/high_mean": 0.0013373524755782757, "clip_ratio/low_mean": 0.0011150982445542468, "clip_ratio/low_min": 0.00030527125175770683, "clip_ratio/region_mean": 0.0024524507229216398, "completions/clipped_ratio": 0.0, "completions/max_length": 666.4, "completions/max_terminated_length": 666.4, "completions/mean_length": 215.2408905029297, "completions/mean_terminated_length": 215.2408905029297, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.029616434887672464, "epoch": 25.686256504438322, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4294758797.0, "reward": 8.033656644821168, "reward_std": 0.34279843121767045, "rewards/combined_reward/mean": 8.033656311035156, "rewards/combined_reward/std": 2.395373451709747, "step": 41970 }, { "clip_ratio/high_max": 0.002236526725755539, "clip_ratio/high_mean": 0.0012537248760054354, "clip_ratio/low_mean": 0.0008072324848399148, "clip_ratio/low_min": 0.00019192206915856028, "clip_ratio/region_mean": 0.002060957354842685, "completions/clipped_ratio": 0.0, "completions/max_length": 585.3, "completions/max_terminated_length": 585.3, "completions/mean_length": 194.04662017822267, "completions/mean_terminated_length": 194.04662017822267, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.02729392599624892, "epoch": 25.70462197734925, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.017822265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4297804096.0, "reward": 7.9741912364959715, "reward_std": 0.3279267311096191, "rewards/combined_reward/mean": 7.974191188812256, "rewards/combined_reward/std": 2.3030303597450255, "step": 42000 }, { "clip_ratio/high_max": 0.0017201717991459494, "clip_ratio/high_mean": 0.00098766798376649, "clip_ratio/low_mean": 0.0007227847539373519, "clip_ratio/low_min": 0.0002250035096949432, "clip_ratio/region_mean": 0.0017104527427970122, "completions/clipped_ratio": 0.0, "completions/max_length": 652.4, "completions/max_terminated_length": 652.4, "completions/mean_length": 201.83151550292968, "completions/mean_terminated_length": 201.83151550292968, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.026454288807387153, "epoch": 25.722375267829815, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.007476806640625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4300879289.0, "reward": 8.090184497833253, "reward_std": 0.2911986202001572, "rewards/combined_reward/mean": 8.090183782577515, "rewards/combined_reward/std": 2.1820542573928834, "step": 42030 }, { "clip_ratio/high_max": 0.001722258629645997, "clip_ratio/high_mean": 0.0009267145230599757, "clip_ratio/low_mean": 0.0007777626376385645, "clip_ratio/low_min": 0.0001881204506692787, "clip_ratio/region_mean": 0.001704477170278551, "completions/clipped_ratio": 0.0, "completions/max_length": 574.3, "completions/max_terminated_length": 574.3, "completions/mean_length": 198.77188110351562, "completions/mean_terminated_length": 198.77188110351562, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.027184322375493744, "epoch": 25.74074074074074, "frac_reward_zero_std": 0.19583334289491178, "grad_norm": 0.005767822265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4303942733.0, "reward": 8.041392278671264, "reward_std": 0.2987912118434906, "rewards/combined_reward/mean": 8.041392040252685, "rewards/combined_reward/std": 2.4485515356063843, "step": 42060 }, { "clip_ratio/high_max": 0.0020153392494345703, "clip_ratio/high_mean": 0.0012072878734519085, "clip_ratio/low_mean": 0.0007153544605292458, "clip_ratio/low_min": 0.00019277768612179595, "clip_ratio/region_mean": 0.0019226423309495052, "completions/clipped_ratio": 0.0, "completions/max_length": 656.7, "completions/max_terminated_length": 656.7, "completions/mean_length": 210.4666732788086, "completions/mean_terminated_length": 210.4666732788086, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.030096099137639006, "epoch": 25.75910621365167, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4307051085.0, "reward": 7.996936941146851, "reward_std": 0.3644949235022068, "rewards/combined_reward/mean": 7.9969367504119875, "rewards/combined_reward/std": 2.415867006778717, "step": 42090 }, { "clip_ratio/high_max": 0.001453043481160421, "clip_ratio/high_mean": 0.0008450083763212509, "clip_ratio/low_mean": 0.0007006899288777883, "clip_ratio/low_min": 0.0001612343852078387, "clip_ratio/region_mean": 0.0015456983232676673, "completions/clipped_ratio": 0.0, "completions/max_length": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 195.19349517822266, "completions/mean_terminated_length": 195.19349517822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.023827529582194983, "epoch": 25.777471686562595, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4310100788.0, "reward": 8.422083568572997, "reward_std": 0.2952167928218842, "rewards/combined_reward/mean": 8.422082996368408, "rewards/combined_reward/std": 1.8900470376014709, "step": 42120 }, { "clip_ratio/high_max": 0.0021065478688493994, "clip_ratio/high_mean": 0.0011631778413478363, "clip_ratio/low_mean": 0.0007086295109426526, "clip_ratio/low_min": 0.0001609973900485784, "clip_ratio/region_mean": 0.0018718073464697228, "completions/clipped_ratio": 0.0, "completions/max_length": 560.8, "completions/max_terminated_length": 560.8, "completions/mean_length": 207.62057647705078, "completions/mean_terminated_length": 207.62057647705078, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.028201520536094905, "epoch": 25.795837159473525, "frac_reward_zero_std": 0.1375000037252903, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4313198211.0, "reward": 8.08601713180542, "reward_std": 0.322512312233448, "rewards/combined_reward/mean": 8.086016941070557, "rewards/combined_reward/std": 2.2964189648628235, "step": 42150 }, { "clip_ratio/high_max": 0.0022384336708152356, "clip_ratio/high_mean": 0.0012795115549427767, "clip_ratio/low_mean": 0.0007055991659096132, "clip_ratio/low_min": 0.0001375601343170274, "clip_ratio/region_mean": 0.001985110718427071, "completions/clipped_ratio": 0.0, "completions/max_length": 665.3, "completions/max_terminated_length": 665.3, "completions/mean_length": 204.82864990234376, "completions/mean_terminated_length": 204.82864990234376, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.03254643927017848, "epoch": 25.81420263238445, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.031494140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4316284913.0, "reward": 7.993771982192993, "reward_std": 0.3106138318777084, "rewards/combined_reward/mean": 7.993771600723266, "rewards/combined_reward/std": 2.176057243347168, "step": 42180 }, { "clip_ratio/high_max": 0.0018702467389327163, "clip_ratio/high_mean": 0.001122829109832916, "clip_ratio/low_mean": 0.0007056361391126605, "clip_ratio/low_min": 0.0002202825014440653, "clip_ratio/region_mean": 0.0018284652636793909, "completions/clipped_ratio": 0.0, "completions/max_length": 548.8, "completions/max_terminated_length": 548.8, "completions/mean_length": 196.72656707763673, "completions/mean_terminated_length": 196.72656707763673, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029128254872436324, "epoch": 25.83256810529538, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.006439208984375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4319340503.0, "reward": 7.787619256973267, "reward_std": 0.3416877567768097, "rewards/combined_reward/mean": 7.787619018554688, "rewards/combined_reward/std": 2.512894105911255, "step": 42210 }, { "clip_ratio/high_max": 0.0015758543318952433, "clip_ratio/high_mean": 0.0008917284860217479, "clip_ratio/low_mean": 0.0006295206600043457, "clip_ratio/low_min": 0.00012228918276377954, "clip_ratio/region_mean": 0.0015212491494215404, "completions/clipped_ratio": 0.0, "completions/max_length": 498.8, "completions/max_terminated_length": 498.8, "completions/mean_length": 193.4612014770508, "completions/mean_terminated_length": 193.4612014770508, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.024124751271059116, "epoch": 25.850933578206305, "frac_reward_zero_std": 0.20000000298023224, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4322383554.0, "reward": 8.177812814712524, "reward_std": 0.2692947812378407, "rewards/combined_reward/mean": 8.177812767028808, "rewards/combined_reward/std": 2.2073087453842164, "step": 42240 }, { "clip_ratio/high_max": 0.002001719011362487, "clip_ratio/high_mean": 0.0011383116093687325, "clip_ratio/low_mean": 0.0006861091568983586, "clip_ratio/low_min": 0.00015144344021488602, "clip_ratio/region_mean": 0.0018244207710570966, "completions/clipped_ratio": 0.0, "completions/max_length": 553.1, "completions/max_terminated_length": 553.1, "completions/mean_length": 197.51953582763673, "completions/mean_terminated_length": 197.51953582763673, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02625508384468655, "epoch": 25.86929905111723, "frac_reward_zero_std": 0.17916667200624942, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4325442189.0, "reward": 8.027224731445312, "reward_std": 0.35890611708164216, "rewards/combined_reward/mean": 8.02722454071045, "rewards/combined_reward/std": 2.4536853551864626, "step": 42270 }, { "clip_ratio/high_max": 0.001734693086958335, "clip_ratio/high_mean": 0.0009703957165280978, "clip_ratio/low_mean": 0.0008482102082780329, "clip_ratio/low_min": 0.0003313726481186071, "clip_ratio/region_mean": 0.001818605930505631, "completions/clipped_ratio": 0.0, "completions/max_length": 592.2, "completions/max_terminated_length": 592.2, "completions/mean_length": 210.29844360351564, "completions/mean_terminated_length": 210.29844360351564, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.031119043954337636, "epoch": 25.88766452402816, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.010498046875, "learning_rate": 5e-06, "loss": 0.003, "num_tokens": 4328549895.0, "reward": 8.000236177444458, "reward_std": 0.4126124709844589, "rewards/combined_reward/mean": 8.00023593902588, "rewards/combined_reward/std": 2.212726855278015, "step": 42300 }, { "clip_ratio/high_max": 0.0017488577907594542, "clip_ratio/high_mean": 0.0009571249639217665, "clip_ratio/low_mean": 0.0008037105736245091, "clip_ratio/low_min": 0.00015965794567212772, "clip_ratio/region_mean": 0.0017608355556149036, "completions/clipped_ratio": 0.0, "completions/max_length": 563.9, "completions/max_terminated_length": 563.9, "completions/mean_length": 203.37630767822264, "completions/mean_terminated_length": 203.37630767822264, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028573476787035663, "epoch": 25.906029996939086, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.007720947265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4331631020.0, "reward": 8.289224672317506, "reward_std": 0.28237667083740237, "rewards/combined_reward/mean": 8.289224338531493, "rewards/combined_reward/std": 1.9990267634391785, "step": 42330 }, { "clip_ratio/high_max": 0.002273830353321197, "clip_ratio/high_mean": 0.00129038371645341, "clip_ratio/low_mean": 0.0007648376859530497, "clip_ratio/low_min": 0.0002197260185009024, "clip_ratio/region_mean": 0.002055221404104183, "completions/clipped_ratio": 0.0, "completions/max_length": 604.3, "completions/max_terminated_length": 604.3, "completions/mean_length": 210.3619857788086, "completions/mean_terminated_length": 210.3619857788086, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.03128150661941618, "epoch": 25.924395469850015, "frac_reward_zero_std": 0.11666666939854622, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4334738970.0, "reward": 7.7500770568847654, "reward_std": 0.3864694982767105, "rewards/combined_reward/mean": 7.750076723098755, "rewards/combined_reward/std": 2.438592517375946, "step": 42360 }, { "clip_ratio/high_max": 0.0018138006756392617, "clip_ratio/high_mean": 0.0010146190235294247, "clip_ratio/low_mean": 0.0007206743424224745, "clip_ratio/low_min": 0.000176781385622841, "clip_ratio/region_mean": 0.0017352933762595057, "completions/clipped_ratio": 0.0, "completions/max_length": 511.4, "completions/max_terminated_length": 511.4, "completions/mean_length": 190.8338607788086, "completions/mean_terminated_length": 190.8338607788086, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.025872059593287606, "epoch": 25.94276094276094, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.0103759765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4337771932.0, "reward": 8.106288862228393, "reward_std": 0.2836348101496696, "rewards/combined_reward/mean": 8.106288385391235, "rewards/combined_reward/std": 1.998074221611023, "step": 42390 }, { "clip_ratio/high_max": 0.0017831996679888106, "clip_ratio/high_mean": 0.001037834167800611, "clip_ratio/low_mean": 0.0007136229370492704, "clip_ratio/low_min": 0.00016504272377157274, "clip_ratio/region_mean": 0.0017514571091548229, "completions/clipped_ratio": 0.0, "completions/max_length": 577.8, "completions/max_terminated_length": 577.8, "completions/mean_length": 216.86380615234376, "completions/mean_terminated_length": 216.86380615234376, "completions/min_length": 51.5, "completions/min_terminated_length": 51.5, "entropy": 0.026589517202228307, "epoch": 25.96112641567187, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.006439208984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4340904849.0, "reward": 7.960796976089478, "reward_std": 0.36680571734905243, "rewards/combined_reward/mean": 7.960796642303467, "rewards/combined_reward/std": 2.370664691925049, "step": 42420 }, { "clip_ratio/high_max": 0.0016309607982596694, "clip_ratio/high_mean": 0.0010331753917853348, "clip_ratio/low_mean": 0.0007699363663050462, "clip_ratio/low_min": 0.0002647808044760798, "clip_ratio/region_mean": 0.0018031117574234183, "completions/clipped_ratio": 0.0, "completions/max_length": 587.6, "completions/max_terminated_length": 587.6, "completions/mean_length": 201.21484985351563, "completions/mean_terminated_length": 201.21484985351563, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.027387155452743173, "epoch": 25.979491888582796, "frac_reward_zero_std": 0.1291666693985462, "grad_norm": 0.0036468505859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4343977674.0, "reward": 8.323296928405762, "reward_std": 0.3406695008277893, "rewards/combined_reward/mean": 8.323296594619752, "rewards/combined_reward/std": 1.899509060382843, "step": 42450 }, { "clip_ratio/high_max": 0.0019064503835882836, "clip_ratio/high_mean": 0.0010415436972834868, "clip_ratio/low_mean": 0.0006306324691953099, "clip_ratio/low_min": 0.0001753962392588922, "clip_ratio/region_mean": 0.0016721761858207174, "completions/clipped_ratio": 0.0, "completions/max_length": 565.2, "completions/max_terminated_length": 565.2, "completions/mean_length": 198.01302490234374, "completions/mean_terminated_length": 198.01302490234374, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.024881696429414055, "epoch": 25.997857361493725, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.00390625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4347038204.0, "reward": 8.004739999771118, "reward_std": 0.3099511802196503, "rewards/combined_reward/mean": 8.004739665985108, "rewards/combined_reward/std": 2.277385103702545, "step": 42480 }, { "clip_ratio/high_max": 0.002029336101497661, "clip_ratio/high_mean": 0.0011859143922891711, "clip_ratio/low_mean": 0.0008333313898383792, "clip_ratio/low_min": 0.0002290940312496921, "clip_ratio/region_mean": 0.00201924578630228, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0909090909091, "completions/max_terminated_length": 647.0909090909091, "completions/mean_length": 206.0800212513317, "completions/mean_terminated_length": 206.0800212513317, "completions/min_length": 41.54545454545455, "completions/min_terminated_length": 41.54545454545455, "entropy": 0.025838665801604262, "epoch": 26.016528925619834, "frac_reward_zero_std": 0.17424242943525314, "grad_norm": 0.0069580078125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4350438862.0, "reward": 8.16370456868952, "reward_std": 0.2898141782392155, "rewards/combined_reward/mean": 8.16370439529419, "rewards/combined_reward/std": 2.1670377146114004, "step": 42510 }, { "clip_ratio/high_max": 0.0023909439201815984, "clip_ratio/high_mean": 0.0014094840682446375, "clip_ratio/low_mean": 0.000880902269394331, "clip_ratio/low_min": 0.00022688326110558894, "clip_ratio/region_mean": 0.002290386309080835, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 206.2960983276367, "completions/mean_terminated_length": 206.2960983276367, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.02780376320394377, "epoch": 26.034894398530763, "frac_reward_zero_std": 0.12916667014360428, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4353531199.0, "reward": 8.120773077011108, "reward_std": 0.40020438134670255, "rewards/combined_reward/mean": 8.12077283859253, "rewards/combined_reward/std": 2.2320034742355346, "step": 42540 }, { "clip_ratio/high_max": 0.0022659020959205616, "clip_ratio/high_mean": 0.001371815483920121, "clip_ratio/low_mean": 0.0008479478319107632, "clip_ratio/low_min": 0.00021610479307128118, "clip_ratio/region_mean": 0.0022197633360822996, "completions/clipped_ratio": 0.0, "completions/max_length": 524.7, "completions/max_terminated_length": 524.7, "completions/mean_length": 198.5065155029297, "completions/mean_terminated_length": 198.5065155029297, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02819079349283129, "epoch": 26.05325987144169, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0120849609375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4356593624.0, "reward": 8.004549026489258, "reward_std": 0.33633205890655515, "rewards/combined_reward/mean": 8.004548788070679, "rewards/combined_reward/std": 2.250439453125, "step": 42570 }, { "clip_ratio/high_max": 0.0026322386009269396, "clip_ratio/high_mean": 0.001377098296385763, "clip_ratio/low_mean": 0.0008334895173902623, "clip_ratio/low_min": 0.0002337305454905921, "clip_ratio/region_mean": 0.0022105877923119504, "completions/clipped_ratio": 0.0, "completions/max_length": 501.3, "completions/max_terminated_length": 501.3, "completions/mean_length": 199.24297332763672, "completions/mean_terminated_length": 199.24297332763672, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02606327013733486, "epoch": 26.07162534435262, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01226806640625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4359658877.0, "reward": 8.283775854110718, "reward_std": 0.25287848263978957, "rewards/combined_reward/mean": 8.28377571105957, "rewards/combined_reward/std": 2.0529332518577577, "step": 42600 }, { "clip_ratio/high_max": 0.0017225668273264697, "clip_ratio/high_mean": 0.0009831109582592035, "clip_ratio/low_mean": 0.0008539601933686451, "clip_ratio/low_min": 0.0002148245386100219, "clip_ratio/region_mean": 0.0018370711449582208, "completions/clipped_ratio": 0.0, "completions/max_length": 552.5, "completions/max_terminated_length": 552.5, "completions/mean_length": 204.05912017822266, "completions/mean_terminated_length": 204.05912017822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02393725796137005, "epoch": 26.089990817263544, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.00506591796875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4362742624.0, "reward": 8.034979343414307, "reward_std": 0.35508296489715574, "rewards/combined_reward/mean": 8.034978866577148, "rewards/combined_reward/std": 2.4079079151153566, "step": 42630 }, { "clip_ratio/high_max": 0.002132286706910236, "clip_ratio/high_mean": 0.0012041705444668575, "clip_ratio/low_mean": 0.0008154718746178939, "clip_ratio/low_min": 0.00024707830998522696, "clip_ratio/region_mean": 0.002019642402107517, "completions/clipped_ratio": 0.0, "completions/max_length": 729.1, "completions/max_terminated_length": 729.1, "completions/mean_length": 212.84479827880858, "completions/mean_terminated_length": 212.84479827880858, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028457320823023716, "epoch": 26.108356290174473, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4365860108.0, "reward": 8.220932626724244, "reward_std": 0.3111332535743713, "rewards/combined_reward/mean": 8.220932340621948, "rewards/combined_reward/std": 2.0954463839530946, "step": 42660 }, { "clip_ratio/high_max": 0.002445998809707817, "clip_ratio/high_mean": 0.0014267963044403587, "clip_ratio/low_mean": 0.0009176247180827583, "clip_ratio/low_min": 0.0002689704507550535, "clip_ratio/region_mean": 0.0023444210198552656, "completions/clipped_ratio": 0.0, "completions/max_length": 619.4, "completions/max_terminated_length": 619.4, "completions/mean_length": 211.2916748046875, "completions/mean_terminated_length": 211.2916748046875, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.029459752685700853, "epoch": 26.1267217630854, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4368971628.0, "reward": 8.122135066986084, "reward_std": 0.3287299364805222, "rewards/combined_reward/mean": 8.122134971618653, "rewards/combined_reward/std": 2.178823393583298, "step": 42690 }, { "clip_ratio/high_max": 0.002300818622946584, "clip_ratio/high_mean": 0.0013385757949436084, "clip_ratio/low_mean": 0.0010045974293461768, "clip_ratio/low_min": 0.000278564969873211, "clip_ratio/region_mean": 0.002343173248906775, "completions/clipped_ratio": 0.0, "completions/max_length": 587.3, "completions/max_terminated_length": 587.3, "completions/mean_length": 202.34037017822266, "completions/mean_terminated_length": 202.34037017822266, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027997578168287872, "epoch": 26.14508723599633, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4372048775.0, "reward": 8.029012870788574, "reward_std": 0.36697239577770235, "rewards/combined_reward/mean": 8.02901258468628, "rewards/combined_reward/std": 2.2164641976356507, "step": 42720 }, { "clip_ratio/high_max": 0.002805983929526216, "clip_ratio/high_mean": 0.0016356730154560258, "clip_ratio/low_mean": 0.000876250961785748, "clip_ratio/low_min": 0.0001992319281271193, "clip_ratio/region_mean": 0.0025119239677830287, "completions/clipped_ratio": 0.0, "completions/max_length": 538.8, "completions/max_terminated_length": 538.8, "completions/mean_length": 205.70990295410155, "completions/mean_terminated_length": 205.70990295410155, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.02936755996197462, "epoch": 26.163452708907254, "frac_reward_zero_std": 0.16250000149011612, "grad_norm": 0.0089111328125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4375138861.0, "reward": 8.27015700340271, "reward_std": 0.3134811371564865, "rewards/combined_reward/mean": 8.270156574249267, "rewards/combined_reward/std": 2.114129602909088, "step": 42750 }, { "clip_ratio/high_max": 0.002036849571353135, "clip_ratio/high_mean": 0.0011748121534765233, "clip_ratio/low_mean": 0.0009116508840331032, "clip_ratio/low_min": 0.00024930368878510006, "clip_ratio/region_mean": 0.0020864630312037963, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 204.47656707763673, "completions/mean_terminated_length": 204.47656707763673, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027036866390456757, "epoch": 26.181818181818183, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4378224211.0, "reward": 8.22464861869812, "reward_std": 0.3109465643763542, "rewards/combined_reward/mean": 8.224648380279541, "rewards/combined_reward/std": 2.0571269154548646, "step": 42780 }, { "clip_ratio/high_max": 0.002730384217284154, "clip_ratio/high_mean": 0.0015020941728532003, "clip_ratio/low_mean": 0.0007866177735801708, "clip_ratio/low_min": 0.0001737684384958508, "clip_ratio/region_mean": 0.0022887119339429773, "completions/clipped_ratio": 0.0, "completions/max_length": 561.4, "completions/max_terminated_length": 561.4, "completions/mean_length": 188.80443267822267, "completions/mean_terminated_length": 188.80443267822267, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028220916679129005, "epoch": 26.20018365472911, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4381249380.0, "reward": 8.16325626373291, "reward_std": 0.2932836651802063, "rewards/combined_reward/mean": 8.163256168365479, "rewards/combined_reward/std": 2.299340617656708, "step": 42810 }, { "clip_ratio/high_max": 0.002334289878247849, "clip_ratio/high_mean": 0.0012306841987689647, "clip_ratio/low_mean": 0.0007643327994931799, "clip_ratio/low_min": 0.00019957243202952668, "clip_ratio/region_mean": 0.001995016996321889, "completions/clipped_ratio": 0.0, "completions/max_length": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 199.69349517822266, "completions/mean_terminated_length": 199.69349517822266, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.024909318098798393, "epoch": 26.21854912764004, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4384316363.0, "reward": 8.067265605926513, "reward_std": 0.30817618817090986, "rewards/combined_reward/mean": 8.06726541519165, "rewards/combined_reward/std": 2.248545730113983, "step": 42840 }, { "clip_ratio/high_max": 0.002141698771447409, "clip_ratio/high_mean": 0.0012083675411001119, "clip_ratio/low_mean": 0.000967021918889562, "clip_ratio/low_min": 0.00029701405001105743, "clip_ratio/region_mean": 0.0021753894824845096, "completions/clipped_ratio": 0.0, "completions/max_length": 550.9, "completions/max_terminated_length": 550.9, "completions/mean_length": 200.79505767822266, "completions/mean_terminated_length": 200.79505767822266, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027368680333408217, "epoch": 26.236914600550964, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4387387576.0, "reward": 7.991748809814453, "reward_std": 0.35030233263969424, "rewards/combined_reward/mean": 7.99174861907959, "rewards/combined_reward/std": 2.2912109732627868, "step": 42870 }, { "clip_ratio/high_max": 0.0023772832360312653, "clip_ratio/high_mean": 0.0012354063687477416, "clip_ratio/low_mean": 0.0009082116071416142, "clip_ratio/low_min": 0.00018104370343886936, "clip_ratio/region_mean": 0.0021436179561230046, "completions/clipped_ratio": 0.0, "completions/max_length": 586.2, "completions/max_terminated_length": 586.2, "completions/mean_length": 200.12630462646484, "completions/mean_terminated_length": 200.12630462646484, "completions/min_length": 47.7, "completions/min_terminated_length": 47.7, "entropy": 0.028066374252860744, "epoch": 26.25528007346189, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4390456221.0, "reward": 8.051167392730713, "reward_std": 0.31835354715585706, "rewards/combined_reward/mean": 8.051167201995849, "rewards/combined_reward/std": 2.192018175125122, "step": 42900 }, { "clip_ratio/high_max": 0.002463647551970401, "clip_ratio/high_mean": 0.0014902541447857705, "clip_ratio/low_mean": 0.0009421484215029825, "clip_ratio/low_min": 0.00028642710919181505, "clip_ratio/region_mean": 0.002432402554162157, "completions/clipped_ratio": 0.0, "completions/max_length": 559.4, "completions/max_terminated_length": 559.4, "completions/mean_length": 201.4367233276367, "completions/mean_terminated_length": 201.4367233276367, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.03163398524435858, "epoch": 26.27364554637282, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4393529898.0, "reward": 8.197664880752564, "reward_std": 0.3471655026078224, "rewards/combined_reward/mean": 8.197664833068847, "rewards/combined_reward/std": 1.998033380508423, "step": 42930 }, { "clip_ratio/high_max": 0.002458932461255851, "clip_ratio/high_mean": 0.0013265714391309303, "clip_ratio/low_mean": 0.0009771469628807002, "clip_ratio/low_min": 0.000214135830901796, "clip_ratio/region_mean": 0.0023037184146232903, "completions/clipped_ratio": 0.0, "completions/max_length": 522.3, "completions/max_terminated_length": 522.3, "completions/mean_length": 193.49297332763672, "completions/mean_terminated_length": 193.49297332763672, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02797096981666982, "epoch": 26.292011019283745, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4396573071.0, "reward": 7.959586238861084, "reward_std": 0.30076816529035566, "rewards/combined_reward/mean": 7.959586048126221, "rewards/combined_reward/std": 2.3803650379180907, "step": 42960 }, { "clip_ratio/high_max": 0.002870865587707764, "clip_ratio/high_mean": 0.001693933039799352, "clip_ratio/low_mean": 0.0011111303674018322, "clip_ratio/low_min": 0.0002674971333666084, "clip_ratio/region_mean": 0.002805063395256487, "completions/clipped_ratio": 0.0, "completions/max_length": 550.8, "completions/max_terminated_length": 550.8, "completions/mean_length": 194.42266235351562, "completions/mean_terminated_length": 194.42266235351562, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.03233007090166211, "epoch": 26.310376492194674, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.006011962890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4399619814.0, "reward": 8.099870014190675, "reward_std": 0.3047137215733528, "rewards/combined_reward/mean": 8.099869728088379, "rewards/combined_reward/std": 2.3322440505027773, "step": 42990 }, { "clip_ratio/high_max": 0.002517221918727349, "clip_ratio/high_mean": 0.0013938236006652005, "clip_ratio/low_mean": 0.0010439418886865799, "clip_ratio/low_min": 0.00025865098359645344, "clip_ratio/region_mean": 0.0024377654612180778, "completions/clipped_ratio": 0.0, "completions/max_length": 600.5, "completions/max_terminated_length": 600.5, "completions/mean_length": 199.4406295776367, "completions/mean_terminated_length": 199.4406295776367, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03187023758267363, "epoch": 26.3287419651056, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4402685826.0, "reward": 8.055952405929565, "reward_std": 0.3451222091913223, "rewards/combined_reward/mean": 8.055952453613282, "rewards/combined_reward/std": 2.226571726799011, "step": 43020 }, { "clip_ratio/high_max": 0.002685547640430741, "clip_ratio/high_mean": 0.001477970638855671, "clip_ratio/low_mean": 0.001031676239169125, "clip_ratio/low_min": 0.00024913659750988397, "clip_ratio/region_mean": 0.002509646881177711, "completions/clipped_ratio": 0.0, "completions/max_length": 557.3, "completions/max_terminated_length": 557.3, "completions/mean_length": 192.1940155029297, "completions/mean_terminated_length": 192.1940155029297, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026512377041702468, "epoch": 26.34710743801653, "frac_reward_zero_std": 0.21666667237877846, "grad_norm": 0.015380859375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4405724011.0, "reward": 8.230492210388183, "reward_std": 0.28661501929163935, "rewards/combined_reward/mean": 8.23049201965332, "rewards/combined_reward/std": 2.129373478889465, "step": 43050 }, { "clip_ratio/high_max": 0.0024015769396404114, "clip_ratio/high_mean": 0.0014216913624598723, "clip_ratio/low_mean": 0.001001877673600878, "clip_ratio/low_min": 0.00035547250566499616, "clip_ratio/region_mean": 0.00242356905437191, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 203.50677642822265, "completions/mean_terminated_length": 203.50677642822265, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02751819205004722, "epoch": 26.365472910927455, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.021728515625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4408805637.0, "reward": 8.301439428329468, "reward_std": 0.30457383692264556, "rewards/combined_reward/mean": 8.301438903808593, "rewards/combined_reward/std": 2.00627943277359, "step": 43080 }, { "clip_ratio/high_max": 0.0022931646434396196, "clip_ratio/high_mean": 0.0013857326775905676, "clip_ratio/low_mean": 0.0009993669107643654, "clip_ratio/low_min": 0.0002616368534897144, "clip_ratio/region_mean": 0.0023850995872635393, "completions/clipped_ratio": 0.0, "completions/max_length": 523.3, "completions/max_terminated_length": 523.3, "completions/mean_length": 196.35599365234376, "completions/mean_terminated_length": 196.35599365234376, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02954672765918076, "epoch": 26.383838383838384, "frac_reward_zero_std": 0.19583334177732467, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4411859804.0, "reward": 8.24140887260437, "reward_std": 0.2643133580684662, "rewards/combined_reward/mean": 8.241408920288086, "rewards/combined_reward/std": 2.0298633933067323, "step": 43110 }, { "clip_ratio/high_max": 0.0020528782471956218, "clip_ratio/high_mean": 0.0011904747603693976, "clip_ratio/low_mean": 0.0009685976419859799, "clip_ratio/low_min": 0.00020895016399056962, "clip_ratio/region_mean": 0.002159072406114622, "completions/clipped_ratio": 0.0, "completions/max_length": 579.9, "completions/max_terminated_length": 579.9, "completions/mean_length": 187.36849365234374, "completions/mean_terminated_length": 187.36849365234374, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.025541597884148358, "epoch": 26.40220385674931, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0125732421875, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 4414879459.0, "reward": 8.291492366790772, "reward_std": 0.22197785526514052, "rewards/combined_reward/mean": 8.291492223739624, "rewards/combined_reward/std": 1.918150818347931, "step": 43140 }, { "clip_ratio/high_max": 0.0022745990238036027, "clip_ratio/high_mean": 0.0013130785362591268, "clip_ratio/low_mean": 0.0009920292298678153, "clip_ratio/low_min": 0.00027421349514042956, "clip_ratio/region_mean": 0.002305107783953038, "completions/clipped_ratio": 0.0, "completions/max_length": 650.9, "completions/max_terminated_length": 650.9, "completions/mean_length": 201.88672332763673, "completions/mean_terminated_length": 201.88672332763673, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.027728812047280373, "epoch": 26.42056932966024, "frac_reward_zero_std": 0.17916667349636556, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4417954864.0, "reward": 8.034603261947632, "reward_std": 0.34207481890916824, "rewards/combined_reward/mean": 8.034603118896484, "rewards/combined_reward/std": 2.277410387992859, "step": 43170 }, { "clip_ratio/high_max": 0.0025884879413448896, "clip_ratio/high_mean": 0.001291240064589753, "clip_ratio/low_mean": 0.0009763220760457141, "clip_ratio/low_min": 0.00023058726656017826, "clip_ratio/region_mean": 0.0022675621413024297, "completions/clipped_ratio": 0.0, "completions/max_length": 507.7, "completions/max_terminated_length": 507.7, "completions/mean_length": 192.7408889770508, "completions/mean_terminated_length": 192.7408889770508, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028145054037061832, "epoch": 26.438934802571165, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4420995149.0, "reward": 7.9157555103302, "reward_std": 0.24335361421108245, "rewards/combined_reward/mean": 7.915755414962769, "rewards/combined_reward/std": 2.589213478565216, "step": 43200 }, { "clip_ratio/high_max": 0.002365837399944818, "clip_ratio/high_mean": 0.001425120670198036, "clip_ratio/low_mean": 0.0010424479927072146, "clip_ratio/low_min": 0.000318248480228552, "clip_ratio/region_mean": 0.0024675686538103035, "completions/clipped_ratio": 0.0, "completions/max_length": 549.1, "completions/max_terminated_length": 549.1, "completions/mean_length": 197.45286865234374, "completions/mean_terminated_length": 197.45286865234374, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.031137637545665106, "epoch": 26.457300275482094, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.0101318359375, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 4424053528.0, "reward": 7.883704900741577, "reward_std": 0.35968592017889023, "rewards/combined_reward/mean": 7.883704757690429, "rewards/combined_reward/std": 2.450859034061432, "step": 43230 }, { "clip_ratio/high_max": 0.0027352115333390734, "clip_ratio/high_mean": 0.00152087604777383, "clip_ratio/low_mean": 0.0010466519392745492, "clip_ratio/low_min": 0.00025556705537989425, "clip_ratio/region_mean": 0.0025675279806212834, "completions/clipped_ratio": 0.0, "completions/max_length": 557.6, "completions/max_terminated_length": 557.6, "completions/mean_length": 198.87474517822267, "completions/mean_terminated_length": 198.87474517822267, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.029693975198703506, "epoch": 26.47566574839302, "frac_reward_zero_std": 0.170833333581686, "grad_norm": 0.01275634765625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4427117367.0, "reward": 8.236129140853881, "reward_std": 0.2453523352742195, "rewards/combined_reward/mean": 8.236129140853881, "rewards/combined_reward/std": 2.2129452347755434, "step": 43260 }, { "clip_ratio/high_max": 0.0023151368969896186, "clip_ratio/high_mean": 0.0012843554873446313, "clip_ratio/low_mean": 0.0009238665532999827, "clip_ratio/low_min": 0.00020318856428881796, "clip_ratio/region_mean": 0.002208222046707912, "completions/clipped_ratio": 0.0, "completions/max_length": 553.3, "completions/max_terminated_length": 553.3, "completions/mean_length": 186.63672332763673, "completions/mean_terminated_length": 186.63672332763673, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.027757968427613376, "epoch": 26.49403122130395, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4430134212.0, "reward": 8.119622325897216, "reward_std": 0.35039139837026595, "rewards/combined_reward/mean": 8.119622087478637, "rewards/combined_reward/std": 2.3529351353645325, "step": 43290 }, { "clip_ratio/high_max": 0.0023408406186111582, "clip_ratio/high_mean": 0.0013048709224676713, "clip_ratio/low_mean": 0.0009390904033352854, "clip_ratio/low_min": 0.00020921610048389993, "clip_ratio/region_mean": 0.00224396133320018, "completions/clipped_ratio": 0.0, "completions/max_length": 606.6, "completions/max_terminated_length": 606.6, "completions/mean_length": 196.96927490234376, "completions/mean_terminated_length": 196.96927490234376, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.029447063663974406, "epoch": 26.512396694214875, "frac_reward_zero_std": 0.1458333395421505, "grad_norm": 0.00885009765625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4433190734.0, "reward": 7.957114934921265, "reward_std": 0.2962640017271042, "rewards/combined_reward/mean": 7.957114791870117, "rewards/combined_reward/std": 2.385941231250763, "step": 43320 }, { "clip_ratio/high_max": 0.0025891436360931646, "clip_ratio/high_mean": 0.0014256027903077968, "clip_ratio/low_mean": 0.0009646395828895038, "clip_ratio/low_min": 0.00026224318159317285, "clip_ratio/region_mean": 0.0023902423767140136, "completions/clipped_ratio": 0.0, "completions/max_length": 556.7, "completions/max_terminated_length": 556.7, "completions/mean_length": 200.21875610351563, "completions/mean_terminated_length": 200.21875610351563, "completions/min_length": 40.7, "completions/min_terminated_length": 40.7, "entropy": 0.03257407893737157, "epoch": 26.530762167125804, "frac_reward_zero_std": 0.21666667237877846, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4436259734.0, "reward": 8.091522455215454, "reward_std": 0.2997145727276802, "rewards/combined_reward/mean": 8.091522169113158, "rewards/combined_reward/std": 2.2314533233642577, "step": 43350 }, { "clip_ratio/high_max": 0.002249197452329099, "clip_ratio/high_mean": 0.0012780054455409603, "clip_ratio/low_mean": 0.0009005079044679102, "clip_ratio/low_min": 0.000283037225259856, "clip_ratio/region_mean": 0.0021785133527979874, "completions/clipped_ratio": 0.0, "completions/max_length": 675.6, "completions/max_terminated_length": 675.6, "completions/mean_length": 203.5221420288086, "completions/mean_terminated_length": 203.5221420288086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02581596150994301, "epoch": 26.54912764003673, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4439341419.0, "reward": 8.10157127380371, "reward_std": 0.289946348965168, "rewards/combined_reward/mean": 8.10157127380371, "rewards/combined_reward/std": 2.2611494302749633, "step": 43380 }, { "clip_ratio/high_max": 0.002696077767662549, "clip_ratio/high_mean": 0.0016034148252704957, "clip_ratio/low_mean": 0.0009401067135210421, "clip_ratio/low_min": 0.00023020950005350944, "clip_ratio/region_mean": 0.0025435215065954253, "completions/clipped_ratio": 0.0, "completions/max_length": 642.3, "completions/max_terminated_length": 642.3, "completions/mean_length": 199.8580780029297, "completions/mean_terminated_length": 199.8580780029297, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.029632346332073213, "epoch": 26.56749311294766, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4442409034.0, "reward": 8.155702447891235, "reward_std": 0.33128407299518586, "rewards/combined_reward/mean": 8.155702161788941, "rewards/combined_reward/std": 1.991120958328247, "step": 43410 }, { "clip_ratio/high_max": 0.0030179434373470335, "clip_ratio/high_mean": 0.0016037358901182111, "clip_ratio/low_mean": 0.0010313768181125246, "clip_ratio/low_min": 0.0003584340905945282, "clip_ratio/region_mean": 0.0026351127264206296, "completions/clipped_ratio": 0.0, "completions/max_length": 578.6, "completions/max_terminated_length": 578.6, "completions/mean_length": 199.87578735351562, "completions/mean_terminated_length": 199.87578735351562, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.030015452376877267, "epoch": 26.585858585858585, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.02001953125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4445476717.0, "reward": 8.11839518547058, "reward_std": 0.3427929446101189, "rewards/combined_reward/mean": 8.118394660949708, "rewards/combined_reward/std": 2.165736138820648, "step": 43440 }, { "clip_ratio/high_max": 0.0022899396256737723, "clip_ratio/high_mean": 0.0012040740915836069, "clip_ratio/low_mean": 0.0008750050679736887, "clip_ratio/low_min": 0.00020696048668469302, "clip_ratio/region_mean": 0.002079079170895663, "completions/clipped_ratio": 0.0, "completions/max_length": 583.7, "completions/max_terminated_length": 583.7, "completions/mean_length": 206.21198272705078, "completions/mean_terminated_length": 206.21198272705078, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02534143125327925, "epoch": 26.604224058769514, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4448568731.0, "reward": 8.496631717681884, "reward_std": 0.27166429311037066, "rewards/combined_reward/mean": 8.496631383895874, "rewards/combined_reward/std": 1.7774954080581664, "step": 43470 }, { "clip_ratio/high_max": 0.002327639395662118, "clip_ratio/high_mean": 0.0012985957832521914, "clip_ratio/low_mean": 0.001002540729799269, "clip_ratio/low_min": 0.00020212642096642715, "clip_ratio/region_mean": 0.0023011364881919388, "completions/clipped_ratio": 0.0, "completions/max_length": 591.5, "completions/max_terminated_length": 591.5, "completions/mean_length": 194.0539108276367, "completions/mean_terminated_length": 194.0539108276367, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.02854435450863093, "epoch": 26.62258953168044, "frac_reward_zero_std": 0.18333333618938924, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4451614058.0, "reward": 7.829999828338623, "reward_std": 0.32055729180574416, "rewards/combined_reward/mean": 7.829999494552612, "rewards/combined_reward/std": 2.4912577152252195, "step": 43500 }, { "clip_ratio/high_max": 0.0022170477042285103, "clip_ratio/high_mean": 0.0011968045820443271, "clip_ratio/low_mean": 0.0010675146272357476, "clip_ratio/low_min": 0.00023226589134234626, "clip_ratio/region_mean": 0.0022643192060058936, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 198.21693115234376, "completions/mean_terminated_length": 198.21693115234376, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.029185956297442316, "epoch": 26.64095500459137, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0152587890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4454675371.0, "reward": 7.950120115280152, "reward_std": 0.3485206201672554, "rewards/combined_reward/mean": 7.950119686126709, "rewards/combined_reward/std": 2.2308396697044373, "step": 43530 }, { "clip_ratio/high_max": 0.0026239483059422733, "clip_ratio/high_mean": 0.0015343068461030876, "clip_ratio/low_mean": 0.001091974622492368, "clip_ratio/low_min": 0.00033951715546815344, "clip_ratio/region_mean": 0.002626281446039987, "completions/clipped_ratio": 0.0, "completions/max_length": 598.8, "completions/max_terminated_length": 598.8, "completions/mean_length": 196.85989990234376, "completions/mean_terminated_length": 196.85989990234376, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 0.028648291155695914, "epoch": 26.659320477502295, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.005279541015625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4457731473.0, "reward": 8.116454029083252, "reward_std": 0.2845679298043251, "rewards/combined_reward/mean": 8.116453456878663, "rewards/combined_reward/std": 2.128688132762909, "step": 43560 }, { "clip_ratio/high_max": 0.0023323865583127676, "clip_ratio/high_mean": 0.0014125217589025852, "clip_ratio/low_mean": 0.0010544963491459687, "clip_ratio/low_min": 0.0002817898564292894, "clip_ratio/region_mean": 0.0024670181068358944, "completions/clipped_ratio": 0.0, "completions/max_length": 682.6, "completions/max_terminated_length": 682.6, "completions/mean_length": 207.4713577270508, "completions/mean_terminated_length": 207.4713577270508, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.02737329315083722, "epoch": 26.677685950413224, "frac_reward_zero_std": 0.16250000111758708, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4460828323.0, "reward": 8.02687110900879, "reward_std": 0.3275641016662121, "rewards/combined_reward/mean": 8.026870632171631, "rewards/combined_reward/std": 2.390163946151733, "step": 43590 }, { "clip_ratio/high_max": 0.0023701276698072133, "clip_ratio/high_mean": 0.0013728518297284609, "clip_ratio/low_mean": 0.0009530760385435618, "clip_ratio/low_min": 0.00032019543117106273, "clip_ratio/region_mean": 0.002325927879185959, "completions/clipped_ratio": 0.0, "completions/max_length": 638.3, "completions/max_terminated_length": 638.3, "completions/mean_length": 204.11068115234374, "completions/mean_terminated_length": 204.11068115234374, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.027993115006635587, "epoch": 26.69605142332415, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4463912268.0, "reward": 7.963768529891968, "reward_std": 0.34070453345775603, "rewards/combined_reward/mean": 7.963768434524536, "rewards/combined_reward/std": 2.3132700324058533, "step": 43620 }, { "clip_ratio/high_max": 0.002872470034829651, "clip_ratio/high_mean": 0.0016612664696973904, "clip_ratio/low_mean": 0.0008524146462756713, "clip_ratio/low_min": 0.00023208850543596782, "clip_ratio/region_mean": 0.002513681115427365, "completions/clipped_ratio": 0.0, "completions/max_length": 715.9, "completions/max_terminated_length": 715.9, "completions/mean_length": 208.16771087646484, "completions/mean_terminated_length": 208.16771087646484, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030044739507138728, "epoch": 26.71441689623508, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.01025390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4467011792.0, "reward": 7.965803098678589, "reward_std": 0.3509227693080902, "rewards/combined_reward/mean": 7.965803050994873, "rewards/combined_reward/std": 2.432839000225067, "step": 43650 }, { "clip_ratio/high_max": 0.002377853580158747, "clip_ratio/high_mean": 0.0012981241077795857, "clip_ratio/low_mean": 0.001116739058852545, "clip_ratio/low_min": 0.00034173508975072765, "clip_ratio/region_mean": 0.0024148631646918753, "completions/clipped_ratio": 0.0, "completions/max_length": 607.4, "completions/max_terminated_length": 607.4, "completions/mean_length": 205.8541717529297, "completions/mean_terminated_length": 205.8541717529297, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.02962331079567472, "epoch": 26.732782369146005, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.01226806640625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4470102432.0, "reward": 7.9489771842956545, "reward_std": 0.38294104784727095, "rewards/combined_reward/mean": 7.948976945877075, "rewards/combined_reward/std": 2.5073593139648436, "step": 43680 }, { "clip_ratio/high_max": 0.002426025346115542, "clip_ratio/high_mean": 0.001351155390511849, "clip_ratio/low_mean": 0.0009407896367823317, "clip_ratio/low_min": 0.0002914871758548543, "clip_ratio/region_mean": 0.00229194502341367, "completions/clipped_ratio": 0.0, "completions/max_length": 563.6, "completions/max_terminated_length": 563.6, "completions/mean_length": 208.07786865234374, "completions/mean_terminated_length": 208.07786865234374, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02892987800296396, "epoch": 26.751147842056934, "frac_reward_zero_std": 0.1375000048428774, "grad_norm": 0.0286865234375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4473201611.0, "reward": 7.897636270523071, "reward_std": 0.3389997027814388, "rewards/combined_reward/mean": 7.897636270523071, "rewards/combined_reward/std": 2.406523251533508, "step": 43710 }, { "clip_ratio/high_max": 0.002990856075242239, "clip_ratio/high_mean": 0.0017671746913644408, "clip_ratio/low_mean": 0.0011240849028884743, "clip_ratio/low_min": 0.00040642768993469266, "clip_ratio/region_mean": 0.002891259607470905, "completions/clipped_ratio": 0.0, "completions/max_length": 612.4, "completions/max_terminated_length": 612.4, "completions/mean_length": 207.32083892822266, "completions/mean_terminated_length": 207.32083892822266, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.034807083569467066, "epoch": 26.76951331496786, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.01251220703125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4476297883.0, "reward": 7.846432638168335, "reward_std": 0.37684765458106995, "rewards/combined_reward/mean": 7.846432399749756, "rewards/combined_reward/std": 2.3885247707366943, "step": 43740 }, { "clip_ratio/high_max": 0.002485616143288401, "clip_ratio/high_mean": 0.0014058345041121356, "clip_ratio/low_mean": 0.0009779731357411948, "clip_ratio/low_min": 0.00022557558719806062, "clip_ratio/region_mean": 0.0023838076517373945, "completions/clipped_ratio": 0.0, "completions/max_length": 623.5, "completions/max_terminated_length": 623.5, "completions/mean_length": 195.13125457763672, "completions/mean_terminated_length": 195.13125457763672, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030560781061649323, "epoch": 26.78787878787879, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4479347347.0, "reward": 7.97910213470459, "reward_std": 0.3238377869129181, "rewards/combined_reward/mean": 7.9791018009185795, "rewards/combined_reward/std": 2.2806135296821592, "step": 43770 }, { "clip_ratio/high_max": 0.002498288616576853, "clip_ratio/high_mean": 0.0013856578264191436, "clip_ratio/low_mean": 0.0011523267523443792, "clip_ratio/low_min": 0.00039389134326484054, "clip_ratio/region_mean": 0.0025379845697898417, "completions/clipped_ratio": 0.0, "completions/max_length": 574.2, "completions/max_terminated_length": 574.2, "completions/mean_length": 203.40234985351563, "completions/mean_terminated_length": 203.40234985351563, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.0321317126819243, "epoch": 26.806244260789715, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4482428572.0, "reward": 7.931872701644897, "reward_std": 0.338101401925087, "rewards/combined_reward/mean": 7.931872606277466, "rewards/combined_reward/std": 2.2972933530807493, "step": 43800 }, { "clip_ratio/high_max": 0.0023493789340136573, "clip_ratio/high_mean": 0.0013859744169167243, "clip_ratio/low_mean": 0.0010315429100349624, "clip_ratio/low_min": 0.00033926873341746006, "clip_ratio/region_mean": 0.002417517332166123, "completions/clipped_ratio": 0.0, "completions/max_length": 586.5, "completions/max_terminated_length": 586.5, "completions/mean_length": 197.9901107788086, "completions/mean_terminated_length": 197.9901107788086, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.029581001633778213, "epoch": 26.824609733700644, "frac_reward_zero_std": 0.17916666865348815, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4485489014.0, "reward": 7.978701972961426, "reward_std": 0.2718758411705494, "rewards/combined_reward/mean": 7.978701877593994, "rewards/combined_reward/std": 2.2396260261535645, "step": 43830 }, { "clip_ratio/high_max": 0.0023995307390578093, "clip_ratio/high_mean": 0.0013841972455945022, "clip_ratio/low_mean": 0.0009589468752892571, "clip_ratio/low_min": 0.0002697788392348836, "clip_ratio/region_mean": 0.002343144104330956, "completions/clipped_ratio": 0.0, "completions/max_length": 527.4, "completions/max_terminated_length": 527.4, "completions/mean_length": 201.17813110351562, "completions/mean_terminated_length": 201.17813110351562, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029209463003401954, "epoch": 26.84297520661157, "frac_reward_zero_std": 0.2000000048428774, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4488561698.0, "reward": 7.773801660537719, "reward_std": 0.4023829013109207, "rewards/combined_reward/mean": 7.773801374435425, "rewards/combined_reward/std": 2.498885786533356, "step": 43860 }, { "clip_ratio/high_max": 0.00227437103894772, "clip_ratio/high_mean": 0.0012635823414408756, "clip_ratio/low_mean": 0.0009312308515291079, "clip_ratio/low_min": 0.0002704313117040632, "clip_ratio/region_mean": 0.0021948131628353926, "completions/clipped_ratio": 0.0, "completions/max_length": 598.6, "completions/max_terminated_length": 598.6, "completions/mean_length": 196.39531860351562, "completions/mean_terminated_length": 196.39531860351562, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.028489110067797203, "epoch": 26.8613406795225, "frac_reward_zero_std": 0.18333334140479565, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4491616016.0, "reward": 8.1447190284729, "reward_std": 0.3561966970562935, "rewards/combined_reward/mean": 8.144718790054322, "rewards/combined_reward/std": 2.1353066086769106, "step": 43890 }, { "clip_ratio/high_max": 0.002580043925748517, "clip_ratio/high_mean": 0.0013531102864362764, "clip_ratio/low_mean": 0.0010238229668175336, "clip_ratio/low_min": 0.0002979639825449946, "clip_ratio/region_mean": 0.0023769332440375973, "completions/clipped_ratio": 0.0, "completions/max_length": 609.6, "completions/max_terminated_length": 609.6, "completions/mean_length": 207.66458892822266, "completions/mean_terminated_length": 207.66458892822266, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.02962183851438264, "epoch": 26.879706152433425, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.01312255859375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4494713608.0, "reward": 7.989966297149659, "reward_std": 0.3201070711016655, "rewards/combined_reward/mean": 7.989966154098511, "rewards/combined_reward/std": 2.2338058590888976, "step": 43920 }, { "clip_ratio/high_max": 0.0024017188411865694, "clip_ratio/high_mean": 0.0013398750658476881, "clip_ratio/low_mean": 0.0009075290291851464, "clip_ratio/low_min": 0.00025481640332145614, "clip_ratio/region_mean": 0.002247404090303462, "completions/clipped_ratio": 0.0, "completions/max_length": 588.2, "completions/max_terminated_length": 588.2, "completions/mean_length": 199.06485137939453, "completions/mean_terminated_length": 199.06485137939453, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.026946257504945, "epoch": 26.898071625344354, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.025146484375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4497778177.0, "reward": 7.724556398391724, "reward_std": 0.3374453917145729, "rewards/combined_reward/mean": 7.7245560646057125, "rewards/combined_reward/std": 2.4353995442390444, "step": 43950 }, { "clip_ratio/high_max": 0.0025109723123023287, "clip_ratio/high_mean": 0.0014334016513506262, "clip_ratio/low_mean": 0.0009411728645015197, "clip_ratio/low_min": 0.00025761911892914214, "clip_ratio/region_mean": 0.0023745745114865714, "completions/clipped_ratio": 0.0, "completions/max_length": 538.7, "completions/max_terminated_length": 538.7, "completions/mean_length": 203.0135467529297, "completions/mean_terminated_length": 203.0135467529297, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028844100159282484, "epoch": 26.91643709825528, "frac_reward_zero_std": 0.18333333507180213, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4500857909.0, "reward": 8.036250257492066, "reward_std": 0.36331894993782043, "rewards/combined_reward/mean": 8.036250019073487, "rewards/combined_reward/std": 2.3510459780693056, "step": 43980 }, { "clip_ratio/high_max": 0.0014358707325300203, "clip_ratio/high_mean": 0.0007955539113027044, "clip_ratio/low_mean": 0.000578896057777456, "clip_ratio/low_min": 9.110370592679828e-05, "clip_ratio/region_mean": 0.0013744499912718311, "completions/clipped_ratio": 0.0, "completions/max_length": 439.25, "completions/max_terminated_length": 439.25, "completions/mean_length": 172.8600311279297, "completions/mean_terminated_length": 172.8600311279297, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.025799620151519775, "epoch": 26.934190388735843, "frac_reward_zero_std": 0.2812500074505806, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4503882543.0, "reward": 8.133254528045654, "reward_std": 0.29021191596984863, "rewards/combined_reward/mean": 8.133254289627075, "rewards/combined_reward/std": 1.8991259634494781, "step": 44010 }, { "clip_ratio/high_max": 0.001729468776708624, "clip_ratio/high_mean": 0.0009151065758487675, "clip_ratio/low_mean": 0.0006791753149930931, "clip_ratio/low_min": 0.00013200593457440846, "clip_ratio/region_mean": 0.0015942818919332544, "completions/clipped_ratio": 0.0, "completions/max_length": 607.3, "completions/max_terminated_length": 607.3, "completions/mean_length": 217.75912170410157, "completions/mean_terminated_length": 217.75912170410157, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.026217846618965267, "epoch": 26.952555861646772, "frac_reward_zero_std": 0.13750000447034835, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4507018898.0, "reward": 8.021451663970947, "reward_std": 0.32156298905611036, "rewards/combined_reward/mean": 8.021451234817505, "rewards/combined_reward/std": 2.202143120765686, "step": 44040 }, { "clip_ratio/high_max": 0.0021780833829931606, "clip_ratio/high_mean": 0.0012793671142086775, "clip_ratio/low_mean": 0.0008537225846036259, "clip_ratio/low_min": 0.0002972242478184247, "clip_ratio/region_mean": 0.002133089699782431, "completions/clipped_ratio": 0.0, "completions/max_length": 565.6, "completions/max_terminated_length": 565.6, "completions/mean_length": 205.45104522705077, "completions/mean_terminated_length": 205.45104522705077, "completions/min_length": 49.2, "completions/min_terminated_length": 49.2, "entropy": 0.02742235961680611, "epoch": 26.970921334557698, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.01434326171875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4510107990.0, "reward": 8.132394552230835, "reward_std": 0.32977318912744524, "rewards/combined_reward/mean": 8.132394170761108, "rewards/combined_reward/std": 2.1116955518722533, "step": 44070 }, { "clip_ratio/high_max": 0.0018109797167805179, "clip_ratio/high_mean": 0.0009756811942982798, "clip_ratio/low_mean": 0.0006289978514663138, "clip_ratio/low_min": 0.00016629105739411897, "clip_ratio/region_mean": 0.001604679050554599, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 200.2510498046875, "completions/mean_terminated_length": 200.2510498046875, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027163404723008475, "epoch": 26.989286807468627, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.03271484375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4513177114.0, "reward": 8.151178121566772, "reward_std": 0.35090481489896774, "rewards/combined_reward/mean": 8.151178073883056, "rewards/combined_reward/std": 2.1257301807403564, "step": 44100 }, { "clip_ratio/high_max": 0.00229615544409827, "clip_ratio/high_mean": 0.0012294603865946735, "clip_ratio/low_mean": 0.0008375712081868217, "clip_ratio/low_min": 0.0002308000016530029, "clip_ratio/region_mean": 0.0020670315969284992, "completions/clipped_ratio": 0.0, "completions/max_length": 556.6, "completions/max_terminated_length": 556.6, "completions/mean_length": 200.9104232788086, "completions/mean_terminated_length": 200.9104232788086, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02677488154502677, "epoch": 27.007958371594736, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4516248770.0, "reward": 7.995694017410278, "reward_std": 0.3561044275760651, "rewards/combined_reward/mean": 7.995693778991699, "rewards/combined_reward/std": 2.2119930386543274, "step": 44130 }, { "clip_ratio/high_max": 0.0023032637070476386, "clip_ratio/high_mean": 0.0012844983568356839, "clip_ratio/low_mean": 0.0009710377593364683, "clip_ratio/low_min": 0.0003021556583310788, "clip_ratio/region_mean": 0.0022555361205983597, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 208.20390777587892, "completions/mean_terminated_length": 208.20390777587892, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.026912606926634906, "epoch": 27.02632384450566, "frac_reward_zero_std": 0.13333333805203437, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4519348433.0, "reward": 8.324887466430663, "reward_std": 0.29589960873126986, "rewards/combined_reward/mean": 8.324886989593505, "rewards/combined_reward/std": 2.0711400628089907, "step": 44160 }, { "clip_ratio/high_max": 0.002295664967581009, "clip_ratio/high_mean": 0.001246698207736093, "clip_ratio/low_mean": 0.000930772860374418, "clip_ratio/low_min": 0.0002678575794561766, "clip_ratio/region_mean": 0.00217747107684166, "completions/clipped_ratio": 0.0, "completions/max_length": 569.2, "completions/max_terminated_length": 569.2, "completions/mean_length": 196.84870300292968, "completions/mean_terminated_length": 196.84870300292968, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.02770581943138192, "epoch": 27.04468931741659, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.0240478515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4522404492.0, "reward": 8.068909072875977, "reward_std": 0.36284712255001067, "rewards/combined_reward/mean": 8.068908882141113, "rewards/combined_reward/std": 2.277169716358185, "step": 44190 }, { "clip_ratio/high_max": 0.002472039267498379, "clip_ratio/high_mean": 0.0014624659310356948, "clip_ratio/low_mean": 0.0009418337502817546, "clip_ratio/low_min": 0.0002631285812337107, "clip_ratio/region_mean": 0.0024042996956268325, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 197.36927490234376, "completions/mean_terminated_length": 197.36927490234376, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02797251963056624, "epoch": 27.063054790327516, "frac_reward_zero_std": 0.17500000707805158, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4525462550.0, "reward": 8.035786724090576, "reward_std": 0.31456571221351626, "rewards/combined_reward/mean": 8.035786390304565, "rewards/combined_reward/std": 2.2895946502685547, "step": 44220 }, { "clip_ratio/high_max": 0.0019505758051915716, "clip_ratio/high_mean": 0.0011189525668063047, "clip_ratio/low_mean": 0.0009874857542551277, "clip_ratio/low_min": 0.00024931689913501034, "clip_ratio/region_mean": 0.0021064383322179007, "completions/clipped_ratio": 0.0, "completions/max_length": 569.8, "completions/max_terminated_length": 569.8, "completions/mean_length": 208.24453582763672, "completions/mean_terminated_length": 208.24453582763672, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.02506372828502208, "epoch": 27.081420263238446, "frac_reward_zero_std": 0.15000000298023225, "grad_norm": 0.006744384765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4528562369.0, "reward": 8.232288026809693, "reward_std": 0.31542485803365705, "rewards/combined_reward/mean": 8.232287979125976, "rewards/combined_reward/std": 2.041542673110962, "step": 44250 }, { "clip_ratio/high_max": 0.0018053815013748438, "clip_ratio/high_mean": 0.001024014536475685, "clip_ratio/low_mean": 0.0007335222561475045, "clip_ratio/low_min": 0.0002174264688316422, "clip_ratio/region_mean": 0.0017575367814667213, "completions/clipped_ratio": 0.0, "completions/max_length": 766.9, "completions/max_terminated_length": 766.9, "completions/mean_length": 208.0057357788086, "completions/mean_terminated_length": 208.0057357788086, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.024545052632068593, "epoch": 27.09978573614937, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4531661271.0, "reward": 8.088837480545044, "reward_std": 0.28710910975933074, "rewards/combined_reward/mean": 8.088837146759033, "rewards/combined_reward/std": 2.4252256035804747, "step": 44280 }, { "clip_ratio/high_max": 0.00250100278450797, "clip_ratio/high_mean": 0.0014174006382139245, "clip_ratio/low_mean": 0.0009330685624566588, "clip_ratio/low_min": 0.00028103169509752967, "clip_ratio/region_mean": 0.0023504691945466522, "completions/clipped_ratio": 0.0, "completions/max_length": 602.1, "completions/max_terminated_length": 602.1, "completions/mean_length": 202.79583740234375, "completions/mean_terminated_length": 202.79583740234375, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.031085442813734213, "epoch": 27.1181512090603, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4534740167.0, "reward": 8.17811107635498, "reward_std": 0.3747496619820595, "rewards/combined_reward/mean": 8.178110504150391, "rewards/combined_reward/std": 2.0442134737968445, "step": 44310 }, { "clip_ratio/high_max": 0.002537441962340381, "clip_ratio/high_mean": 0.0014935655860729943, "clip_ratio/low_mean": 0.001027983848932005, "clip_ratio/low_min": 0.00032559858858197306, "clip_ratio/region_mean": 0.0025215494310638557, "completions/clipped_ratio": 0.0, "completions/max_length": 629.5, "completions/max_terminated_length": 629.5, "completions/mean_length": 210.13880920410156, "completions/mean_terminated_length": 210.13880920410156, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.02924740343199422, "epoch": 27.136516681971226, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 4537847260.0, "reward": 8.153267335891723, "reward_std": 0.25735995918512344, "rewards/combined_reward/mean": 8.153266859054565, "rewards/combined_reward/std": 2.2084373831748962, "step": 44340 }, { "clip_ratio/high_max": 0.0026201508842253435, "clip_ratio/high_mean": 0.0015122217206226195, "clip_ratio/low_mean": 0.0010419162010900133, "clip_ratio/low_min": 0.00034483524408036225, "clip_ratio/region_mean": 0.002554137905826792, "completions/clipped_ratio": 0.0, "completions/max_length": 609.2, "completions/max_terminated_length": 609.2, "completions/mean_length": 202.97656860351563, "completions/mean_terminated_length": 202.97656860351563, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.030042209189074735, "epoch": 27.154882154882156, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4540926850.0, "reward": 8.118765354156494, "reward_std": 0.34228516519069674, "rewards/combined_reward/mean": 8.118764925003052, "rewards/combined_reward/std": 2.0475998163223266, "step": 44370 }, { "clip_ratio/high_max": 0.0022482299185260977, "clip_ratio/high_mean": 0.001234381925435931, "clip_ratio/low_mean": 0.000795054985974275, "clip_ratio/low_min": 0.0002474743091928152, "clip_ratio/region_mean": 0.002029436904316147, "completions/clipped_ratio": 0.0, "completions/max_length": 564.2, "completions/max_terminated_length": 564.2, "completions/mean_length": 209.8338607788086, "completions/mean_terminated_length": 209.8338607788086, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.026435402082279326, "epoch": 27.17324762779308, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4544032772.0, "reward": 8.219521808624268, "reward_std": 0.3193818345665932, "rewards/combined_reward/mean": 8.219521617889404, "rewards/combined_reward/std": 2.1362537860870363, "step": 44400 }, { "clip_ratio/high_max": 0.0023122149335298068, "clip_ratio/high_mean": 0.001376952790208937, "clip_ratio/low_mean": 0.0010201622394864292, "clip_ratio/low_min": 0.00024502278635433566, "clip_ratio/region_mean": 0.002397115019266494, "completions/clipped_ratio": 0.0, "completions/max_length": 541.7, "completions/max_terminated_length": 541.7, "completions/mean_length": 196.08672485351562, "completions/mean_terminated_length": 196.08672485351562, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.029522808610151212, "epoch": 27.19161310070401, "frac_reward_zero_std": 0.1333333358168602, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4547085905.0, "reward": 8.186588764190674, "reward_std": 0.2569658517837524, "rewards/combined_reward/mean": 8.186588382720947, "rewards/combined_reward/std": 2.3090619444847107, "step": 44430 }, { "clip_ratio/high_max": 0.0023099714497220704, "clip_ratio/high_mean": 0.0013111224619933638, "clip_ratio/low_mean": 0.0007899223089528581, "clip_ratio/low_min": 0.0001987585441383999, "clip_ratio/region_mean": 0.0021010447739778708, "completions/clipped_ratio": 0.0, "completions/max_length": 573.7, "completions/max_terminated_length": 573.7, "completions/mean_length": 194.1822952270508, "completions/mean_terminated_length": 194.1822952270508, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.02672280164745947, "epoch": 27.209978573614936, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4550131725.0, "reward": 8.19028344154358, "reward_std": 0.2697788283228874, "rewards/combined_reward/mean": 8.190283250808715, "rewards/combined_reward/std": 2.1358622789382933, "step": 44460 }, { "clip_ratio/high_max": 0.002293652425093266, "clip_ratio/high_mean": 0.0012708114538933538, "clip_ratio/low_mean": 0.0007508089510035158, "clip_ratio/low_min": 0.00015340108366217463, "clip_ratio/region_mean": 0.0020216204318179126, "completions/clipped_ratio": 0.0, "completions/max_length": 678.6, "completions/max_terminated_length": 678.6, "completions/mean_length": 199.0992233276367, "completions/mean_terminated_length": 199.0992233276367, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.025965992299218973, "epoch": 27.228344046525866, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.0150146484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4553196426.0, "reward": 7.978412437438965, "reward_std": 0.37085999101400374, "rewards/combined_reward/mean": 7.978412246704101, "rewards/combined_reward/std": 2.3510880708694457, "step": 44490 }, { "clip_ratio/high_max": 0.0021328167570269824, "clip_ratio/high_mean": 0.0012021783684758702, "clip_ratio/low_mean": 0.0009048887430253671, "clip_ratio/low_min": 0.00027274784151813946, "clip_ratio/region_mean": 0.002107067106893131, "completions/clipped_ratio": 0.0, "completions/max_length": 559.1, "completions/max_terminated_length": 559.1, "completions/mean_length": 196.99844207763672, "completions/mean_terminated_length": 196.99844207763672, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.026364575955085456, "epoch": 27.24670951943679, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4556253060.0, "reward": 8.05042905807495, "reward_std": 0.3474843680858612, "rewards/combined_reward/mean": 8.050428581237792, "rewards/combined_reward/std": 2.2318315625190737, "step": 44520 }, { "clip_ratio/high_max": 0.0030754089006222785, "clip_ratio/high_mean": 0.001566407081312112, "clip_ratio/low_mean": 0.0008775858825780839, "clip_ratio/low_min": 0.00021606622879820255, "clip_ratio/region_mean": 0.00244399297516793, "completions/clipped_ratio": 0.0, "completions/max_length": 579.3, "completions/max_terminated_length": 579.3, "completions/mean_length": 197.5901092529297, "completions/mean_terminated_length": 197.5901092529297, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03305668981435398, "epoch": 27.26507499234772, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 1.78125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4559311966.0, "reward": 8.148471117019653, "reward_std": 0.3955563485622406, "rewards/combined_reward/mean": 8.14847092628479, "rewards/combined_reward/std": 2.1132936477661133, "step": 44550 }, { "clip_ratio/high_max": 0.0027848483878187835, "clip_ratio/high_mean": 0.0015762485874195895, "clip_ratio/low_mean": 0.0010185174320213263, "clip_ratio/low_min": 0.0003513524791439219, "clip_ratio/region_mean": 0.0025947660299910544, "completions/clipped_ratio": 0.0, "completions/max_length": 585.8, "completions/max_terminated_length": 585.8, "completions/mean_length": 200.0968811035156, "completions/mean_terminated_length": 200.0968811035156, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.026677497119332354, "epoch": 27.283440465258646, "frac_reward_zero_std": 0.22500000298023223, "grad_norm": 0.00701904296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4562380498.0, "reward": 8.081832695007325, "reward_std": 0.2778986021876335, "rewards/combined_reward/mean": 8.08183250427246, "rewards/combined_reward/std": 2.154601287841797, "step": 44580 }, { "clip_ratio/high_max": 0.0023883246326780257, "clip_ratio/high_mean": 0.001344413288946574, "clip_ratio/low_mean": 0.0008628346912397926, "clip_ratio/low_min": 0.00022054719496130323, "clip_ratio/region_mean": 0.002207247988553718, "completions/clipped_ratio": 0.0, "completions/max_length": 515.8, "completions/max_terminated_length": 515.8, "completions/mean_length": 203.0869873046875, "completions/mean_terminated_length": 203.0869873046875, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.027548167342320084, "epoch": 27.301805938169576, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4565460512.0, "reward": 8.045514583587646, "reward_std": 0.33242527171969416, "rewards/combined_reward/mean": 8.04551429748535, "rewards/combined_reward/std": 2.347574031352997, "step": 44610 }, { "clip_ratio/high_max": 0.0029496087702379254, "clip_ratio/high_mean": 0.0017188939050053402, "clip_ratio/low_mean": 0.0009227464462431574, "clip_ratio/low_min": 0.00022610896436769205, "clip_ratio/region_mean": 0.002641640356159769, "completions/clipped_ratio": 0.0, "completions/max_length": 656.3, "completions/max_terminated_length": 656.3, "completions/mean_length": 193.7257858276367, "completions/mean_terminated_length": 193.7257858276367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.036720190197229385, "epoch": 27.3201714110805, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.00335693359375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4568504579.0, "reward": 8.021978664398194, "reward_std": 0.2797660008072853, "rewards/combined_reward/mean": 8.021978569030761, "rewards/combined_reward/std": 2.366086745262146, "step": 44640 }, { "clip_ratio/high_max": 0.0022463980121149993, "clip_ratio/high_mean": 0.001290455301204929, "clip_ratio/low_mean": 0.0010173074017075124, "clip_ratio/low_min": 0.0002686551415536087, "clip_ratio/region_mean": 0.0023077627299547504, "completions/clipped_ratio": 0.0, "completions/max_length": 511.1, "completions/max_terminated_length": 511.1, "completions/mean_length": 193.15677642822266, "completions/mean_terminated_length": 193.15677642822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02754853186197579, "epoch": 27.33853688399143, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.004364013671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4571546461.0, "reward": 8.243097305297852, "reward_std": 0.2742438532412052, "rewards/combined_reward/mean": 8.243096923828125, "rewards/combined_reward/std": 1.9933247804641723, "step": 44670 }, { "clip_ratio/high_max": 0.002683098587052276, "clip_ratio/high_mean": 0.001486658138067772, "clip_ratio/low_mean": 0.0009798990186633697, "clip_ratio/low_min": 0.00025021934391891894, "clip_ratio/region_mean": 0.002466557159884057, "completions/clipped_ratio": 0.0, "completions/max_length": 686.3, "completions/max_terminated_length": 686.3, "completions/mean_length": 210.4143264770508, "completions/mean_terminated_length": 210.4143264770508, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.029752965737134218, "epoch": 27.356902356902356, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4574654612.0, "reward": 8.221633052825927, "reward_std": 0.2805840909481049, "rewards/combined_reward/mean": 8.221632862091065, "rewards/combined_reward/std": 2.0694217801094057, "step": 44700 }, { "clip_ratio/high_max": 0.0028843754926735224, "clip_ratio/high_mean": 0.0015726997698948253, "clip_ratio/low_mean": 0.0009376723506041647, "clip_ratio/low_min": 0.00021103391991346144, "clip_ratio/region_mean": 0.0025103721340807776, "completions/clipped_ratio": 0.0, "completions/max_length": 552.5, "completions/max_terminated_length": 552.5, "completions/mean_length": 196.7127670288086, "completions/mean_terminated_length": 196.7127670288086, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.029646611671584347, "epoch": 27.375267829813286, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4577710149.0, "reward": 8.216882038116456, "reward_std": 0.2864066958427429, "rewards/combined_reward/mean": 8.216881895065308, "rewards/combined_reward/std": 2.155874562263489, "step": 44730 }, { "clip_ratio/high_max": 0.002308606362445668, "clip_ratio/high_mean": 0.0013022510329998719, "clip_ratio/low_mean": 0.0009073702639094942, "clip_ratio/low_min": 0.00019670880428748205, "clip_ratio/region_mean": 0.0022096213113400156, "completions/clipped_ratio": 0.0, "completions/max_length": 479.3, "completions/max_terminated_length": 479.3, "completions/mean_length": 189.30443420410157, "completions/mean_terminated_length": 189.30443420410157, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.026067022358377774, "epoch": 27.39363330272421, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4580737238.0, "reward": 8.34040446281433, "reward_std": 0.24227860569953918, "rewards/combined_reward/mean": 8.340404081344605, "rewards/combined_reward/std": 1.9457952618598937, "step": 44760 }, { "clip_ratio/high_max": 0.002317119472233268, "clip_ratio/high_mean": 0.0012871062909349954, "clip_ratio/low_mean": 0.0008257243596744956, "clip_ratio/low_min": 0.00020374711651432637, "clip_ratio/region_mean": 0.002112830656308991, "completions/clipped_ratio": 0.0, "completions/max_length": 641.7, "completions/max_terminated_length": 641.7, "completions/mean_length": 191.3585998535156, "completions/mean_terminated_length": 191.3585998535156, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.02602783207160731, "epoch": 27.41199877563514, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4583772215.0, "reward": 8.114229774475097, "reward_std": 0.29815505594015124, "rewards/combined_reward/mean": 8.114229679107666, "rewards/combined_reward/std": 2.0393327355384825, "step": 44790 }, { "clip_ratio/high_max": 0.002404391373662899, "clip_ratio/high_mean": 0.0013808638277017357, "clip_ratio/low_mean": 0.000988934685907831, "clip_ratio/low_min": 0.00027623513039240304, "clip_ratio/region_mean": 0.0023697985016042367, "completions/clipped_ratio": 0.0, "completions/max_length": 671.9, "completions/max_terminated_length": 671.9, "completions/mean_length": 204.5572967529297, "completions/mean_terminated_length": 204.5572967529297, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.029068869748152794, "epoch": 27.430364248546066, "frac_reward_zero_std": 0.18750000856816768, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4586857875.0, "reward": 8.054925298690796, "reward_std": 0.2865504786372185, "rewards/combined_reward/mean": 8.054925060272216, "rewards/combined_reward/std": 2.4836872100830076, "step": 44820 }, { "clip_ratio/high_max": 0.002561680878473756, "clip_ratio/high_mean": 0.0014029721668824399, "clip_ratio/low_mean": 0.0008606110022810753, "clip_ratio/low_min": 0.00031591965259091616, "clip_ratio/region_mean": 0.0022635831856556857, "completions/clipped_ratio": 0.0, "completions/max_length": 498.3, "completions/max_terminated_length": 498.3, "completions/mean_length": 191.26693267822264, "completions/mean_terminated_length": 191.26693267822264, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028555132386585077, "epoch": 27.448729721456996, "frac_reward_zero_std": 0.1916666690260172, "grad_norm": 0.0101318359375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4589892500.0, "reward": 7.830395269393921, "reward_std": 0.3792285233736038, "rewards/combined_reward/mean": 7.830395078659057, "rewards/combined_reward/std": 2.542549455165863, "step": 44850 }, { "clip_ratio/high_max": 0.002201139101816807, "clip_ratio/high_mean": 0.0013334753394398527, "clip_ratio/low_mean": 0.0009677355121918178, "clip_ratio/low_min": 0.00032226193264553633, "clip_ratio/region_mean": 0.0023012108576949686, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 203.01536712646484, "completions/mean_terminated_length": 203.01536712646484, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.027833868757200737, "epoch": 27.46709519436792, "frac_reward_zero_std": 0.15000000409781933, "grad_norm": 0.01092529296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4592972239.0, "reward": 8.086072397232055, "reward_std": 0.2920348420739174, "rewards/combined_reward/mean": 8.086071920394897, "rewards/combined_reward/std": 2.332474875450134, "step": 44880 }, { "clip_ratio/high_max": 0.002806774694666577, "clip_ratio/high_mean": 0.0015723309371727132, "clip_ratio/low_mean": 0.001111783695826792, "clip_ratio/low_min": 0.00026808340238252033, "clip_ratio/region_mean": 0.002684114608564414, "completions/clipped_ratio": 0.0, "completions/max_length": 540.7, "completions/max_terminated_length": 540.7, "completions/mean_length": 194.83646545410156, "completions/mean_terminated_length": 194.83646545410156, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.031129319593310355, "epoch": 27.48546066727885, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4596020571.0, "reward": 8.19278268814087, "reward_std": 0.3187647297978401, "rewards/combined_reward/mean": 8.192782258987426, "rewards/combined_reward/std": 2.1819283843040465, "step": 44910 }, { "clip_ratio/high_max": 0.0020880787041581547, "clip_ratio/high_mean": 0.0011287554676528088, "clip_ratio/low_mean": 0.0008178046064131195, "clip_ratio/low_min": 0.00018745333218248562, "clip_ratio/region_mean": 0.0019465600552697046, "completions/clipped_ratio": 0.0, "completions/max_length": 612.8, "completions/max_terminated_length": 612.8, "completions/mean_length": 195.18151397705077, "completions/mean_terminated_length": 195.18151397705077, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.024772489489987493, "epoch": 27.503826140189776, "frac_reward_zero_std": 0.18750000409781933, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4599070228.0, "reward": 8.029117727279663, "reward_std": 0.3374523460865021, "rewards/combined_reward/mean": 8.029117155075074, "rewards/combined_reward/std": 2.376683938503265, "step": 44940 }, { "clip_ratio/high_max": 0.0028367875202093273, "clip_ratio/high_mean": 0.001597776474773127, "clip_ratio/low_mean": 0.0011255640398303512, "clip_ratio/low_min": 0.0002769372906186618, "clip_ratio/region_mean": 0.0027233404898045896, "completions/clipped_ratio": 0.0, "completions/max_length": 646.6, "completions/max_terminated_length": 646.6, "completions/mean_length": 198.1812545776367, "completions/mean_terminated_length": 198.1812545776367, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03603890573916336, "epoch": 27.522191613100706, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.004608154296875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4602131404.0, "reward": 7.803977918624878, "reward_std": 0.27429828643798826, "rewards/combined_reward/mean": 7.803977680206299, "rewards/combined_reward/std": 2.6392765760421755, "step": 44970 }, { "clip_ratio/high_max": 0.0017346713235989835, "clip_ratio/high_mean": 0.0010392355239067304, "clip_ratio/low_mean": 0.0009322992577532811, "clip_ratio/low_min": 0.0002790415494625146, "clip_ratio/region_mean": 0.0019715347919069853, "completions/clipped_ratio": 0.0, "completions/max_length": 669.3, "completions/max_terminated_length": 669.3, "completions/mean_length": 202.45547332763672, "completions/mean_terminated_length": 202.45547332763672, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.025887468938405316, "epoch": 27.54055708601163, "frac_reward_zero_std": 0.18750000670552253, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4605208993.0, "reward": 8.418243980407714, "reward_std": 0.2870872303843498, "rewards/combined_reward/mean": 8.418244028091431, "rewards/combined_reward/std": 1.822973644733429, "step": 45000 }, { "clip_ratio/high_max": 0.002635974004806485, "clip_ratio/high_mean": 0.0014956036393414252, "clip_ratio/low_mean": 0.0009951203036810815, "clip_ratio/low_min": 0.00021809233076055534, "clip_ratio/region_mean": 0.0024907239402333897, "completions/clipped_ratio": 0.0, "completions/max_length": 626.7, "completions/max_terminated_length": 626.7, "completions/mean_length": 202.59844207763672, "completions/mean_terminated_length": 202.59844207763672, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.028652083796138563, "epoch": 27.55892255892256, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4608287131.0, "reward": 8.134448528289795, "reward_std": 0.3202049255371094, "rewards/combined_reward/mean": 8.134448289871216, "rewards/combined_reward/std": 2.156070375442505, "step": 45030 }, { "clip_ratio/high_max": 0.002782245799122999, "clip_ratio/high_mean": 0.0016406096307036932, "clip_ratio/low_mean": 0.0009303968660484921, "clip_ratio/low_min": 0.0002512529094625885, "clip_ratio/region_mean": 0.0025710064578258122, "completions/clipped_ratio": 0.0, "completions/max_length": 621.9, "completions/max_terminated_length": 621.9, "completions/mean_length": 203.3151107788086, "completions/mean_terminated_length": 203.3151107788086, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029944360007842383, "epoch": 27.577288031833486, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.00732421875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4611368021.0, "reward": 7.990884923934937, "reward_std": 0.3540525421500206, "rewards/combined_reward/mean": 7.990884637832641, "rewards/combined_reward/std": 2.1613080978393553, "step": 45060 }, { "clip_ratio/high_max": 0.0027558831396163443, "clip_ratio/high_mean": 0.00149124095866379, "clip_ratio/low_mean": 0.0007091726747603389, "clip_ratio/low_min": 0.00019635370093358992, "clip_ratio/region_mean": 0.0022004136219038626, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 200.61927185058593, "completions/mean_terminated_length": 200.61927185058593, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.027427873318083586, "epoch": 27.595653504744412, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4614438559.0, "reward": 8.33738398551941, "reward_std": 0.32771536707878113, "rewards/combined_reward/mean": 8.33738398551941, "rewards/combined_reward/std": 2.11200088262558, "step": 45090 }, { "clip_ratio/high_max": 0.0022630583218415266, "clip_ratio/high_mean": 0.0012678371195458263, "clip_ratio/low_mean": 0.0007666268585429255, "clip_ratio/low_min": 0.00015274976370468116, "clip_ratio/region_mean": 0.0020344639725711506, "completions/clipped_ratio": 0.0, "completions/max_length": 586.7, "completions/max_terminated_length": 586.7, "completions/mean_length": 204.8033889770508, "completions/mean_terminated_length": 204.8033889770508, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.025022524626304707, "epoch": 27.61401897765534, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4617525164.0, "reward": 8.235228872299194, "reward_std": 0.30814793407917024, "rewards/combined_reward/mean": 8.235228681564331, "rewards/combined_reward/std": 2.0018701195716857, "step": 45120 }, { "clip_ratio/high_max": 0.0027524930279469117, "clip_ratio/high_mean": 0.0015350697412941372, "clip_ratio/low_mean": 0.0010617953071535642, "clip_ratio/low_min": 0.00032792097529939684, "clip_ratio/region_mean": 0.0025968650278324883, "completions/clipped_ratio": 0.0, "completions/max_length": 517.9, "completions/max_terminated_length": 517.9, "completions/mean_length": 191.13906707763672, "completions/mean_terminated_length": 191.13906707763672, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.032105647431065636, "epoch": 27.63238445056627, "frac_reward_zero_std": 0.2125000052154064, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4620559298.0, "reward": 7.727914237976075, "reward_std": 0.31710233688354494, "rewards/combined_reward/mean": 7.727914190292358, "rewards/combined_reward/std": 2.4234054327011108, "step": 45150 }, { "clip_ratio/high_max": 0.0018912923042080365, "clip_ratio/high_mean": 0.0011042257475007014, "clip_ratio/low_mean": 0.0009155450769564292, "clip_ratio/low_min": 0.00024457409187258843, "clip_ratio/region_mean": 0.0020197708441022162, "completions/clipped_ratio": 0.0, "completions/max_length": 573.7, "completions/max_terminated_length": 573.7, "completions/mean_length": 205.1252670288086, "completions/mean_terminated_length": 205.1252670288086, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.027790050332744915, "epoch": 27.650749923477196, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4623647139.0, "reward": 8.040917539596558, "reward_std": 0.3142738237977028, "rewards/combined_reward/mean": 8.040917253494262, "rewards/combined_reward/std": 2.2673136711120607, "step": 45180 }, { "clip_ratio/high_max": 0.002625857708820452, "clip_ratio/high_mean": 0.0014545263921415122, "clip_ratio/low_mean": 0.000999250542857529, "clip_ratio/low_min": 0.0003078384320057618, "clip_ratio/region_mean": 0.002453776957311978, "completions/clipped_ratio": 0.0, "completions/max_length": 617.8, "completions/max_terminated_length": 617.8, "completions/mean_length": 197.91510925292968, "completions/mean_terminated_length": 197.91510925292968, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.028744263201951982, "epoch": 27.669115396388122, "frac_reward_zero_std": 0.17500000596046447, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4626707293.0, "reward": 8.161781215667725, "reward_std": 0.259431329369545, "rewards/combined_reward/mean": 8.161781072616577, "rewards/combined_reward/std": 2.0749451875686646, "step": 45210 }, { "clip_ratio/high_max": 0.002368820055077473, "clip_ratio/high_mean": 0.0014253490212771188, "clip_ratio/low_mean": 0.0009058865170421389, "clip_ratio/low_min": 0.00025880898804947114, "clip_ratio/region_mean": 0.0023312355440187576, "completions/clipped_ratio": 0.0, "completions/max_length": 643.5, "completions/max_terminated_length": 643.5, "completions/mean_length": 211.5914093017578, "completions/mean_terminated_length": 211.5914093017578, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028925464550654093, "epoch": 27.68748086929905, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.0228271484375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4629819964.0, "reward": 7.956461620330811, "reward_std": 0.35910646319389344, "rewards/combined_reward/mean": 7.956461429595947, "rewards/combined_reward/std": 2.55367693901062, "step": 45240 }, { "clip_ratio/high_max": 0.002219088122365065, "clip_ratio/high_mean": 0.0011983336650397784, "clip_ratio/low_mean": 0.0009368891132908175, "clip_ratio/low_min": 0.0002369027604193737, "clip_ratio/region_mean": 0.0021352227835450322, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 196.72656860351563, "completions/mean_terminated_length": 196.72656860351563, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.026645007729530334, "epoch": 27.705846342209977, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4632875554.0, "reward": 7.96364197731018, "reward_std": 0.3370669215917587, "rewards/combined_reward/mean": 7.963641595840454, "rewards/combined_reward/std": 2.251873052120209, "step": 45270 }, { "clip_ratio/high_max": 0.0023006354400422423, "clip_ratio/high_mean": 0.0013709632046811748, "clip_ratio/low_mean": 0.0010104410592854643, "clip_ratio/low_min": 0.00026265318204726404, "clip_ratio/region_mean": 0.0023814042855519802, "completions/clipped_ratio": 0.0, "completions/max_length": 616.8, "completions/max_terminated_length": 616.8, "completions/mean_length": 210.9445358276367, "completions/mean_terminated_length": 210.9445358276367, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.028206066647544503, "epoch": 27.724211815120906, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4635985741.0, "reward": 8.082234239578247, "reward_std": 0.35085093379020693, "rewards/combined_reward/mean": 8.082233905792236, "rewards/combined_reward/std": 2.422307574748993, "step": 45300 }, { "clip_ratio/high_max": 0.0023954149151298528, "clip_ratio/high_mean": 0.0014021922139363595, "clip_ratio/low_mean": 0.0010512380157403336, "clip_ratio/low_min": 0.000331398186244769, "clip_ratio/region_mean": 0.002453430193903235, "completions/clipped_ratio": 0.0, "completions/max_length": 579.6, "completions/max_terminated_length": 579.6, "completions/mean_length": 205.26510925292968, "completions/mean_terminated_length": 205.26510925292968, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.029867597358922162, "epoch": 27.742577288031832, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0064697265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4639074119.0, "reward": 7.853780269622803, "reward_std": 0.34617658853530886, "rewards/combined_reward/mean": 7.853780078887939, "rewards/combined_reward/std": 2.579101026058197, "step": 45330 }, { "clip_ratio/high_max": 0.002758355692397648, "clip_ratio/high_mean": 0.0016576676867164982, "clip_ratio/low_mean": 0.0010782310587577133, "clip_ratio/low_min": 0.00036409870347900626, "clip_ratio/region_mean": 0.0027358987552967545, "completions/clipped_ratio": 0.0, "completions/max_length": 558.9, "completions/max_terminated_length": 558.9, "completions/mean_length": 213.77448272705078, "completions/mean_terminated_length": 213.77448272705078, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.03186916057020426, "epoch": 27.76094276094276, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4642195173.0, "reward": 7.986157512664795, "reward_std": 0.37749162018299104, "rewards/combined_reward/mean": 7.986157369613648, "rewards/combined_reward/std": 2.2298384308815002, "step": 45360 }, { "clip_ratio/high_max": 0.0029863049766087595, "clip_ratio/high_mean": 0.0017623049139122789, "clip_ratio/low_mean": 0.0012768183234584285, "clip_ratio/low_min": 0.0005630364646397842, "clip_ratio/region_mean": 0.00303912321687676, "completions/clipped_ratio": 0.0, "completions/max_length": 586.1, "completions/max_terminated_length": 586.1, "completions/mean_length": 208.30599517822264, "completions/mean_terminated_length": 208.30599517822264, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.03557104294498761, "epoch": 27.779308233853687, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4645295228.0, "reward": 7.758494567871094, "reward_std": 0.3573801785707474, "rewards/combined_reward/mean": 7.758494520187378, "rewards/combined_reward/std": 2.504550504684448, "step": 45390 }, { "clip_ratio/high_max": 0.002774492530928304, "clip_ratio/high_mean": 0.001471383628207453, "clip_ratio/low_mean": 0.0010663783005535757, "clip_ratio/low_min": 0.0003334427710797172, "clip_ratio/region_mean": 0.002537761933247869, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 191.16901397705078, "completions/mean_terminated_length": 191.16901397705078, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.0310110443426917, "epoch": 27.797673706764616, "frac_reward_zero_std": 0.2041666716337204, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4648329477.0, "reward": 7.872711229324341, "reward_std": 0.3343114942312241, "rewards/combined_reward/mean": 7.8727109909057615, "rewards/combined_reward/std": 2.2811044096946715, "step": 45420 }, { "clip_ratio/high_max": 0.002288023580331355, "clip_ratio/high_mean": 0.001397400690742264, "clip_ratio/low_mean": 0.0010810225268263216, "clip_ratio/low_min": 0.0003047392667212989, "clip_ratio/region_mean": 0.002478423200470085, "completions/clipped_ratio": 0.0, "completions/max_length": 602.2, "completions/max_terminated_length": 602.2, "completions/mean_length": 201.6315155029297, "completions/mean_terminated_length": 201.6315155029297, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.030595695630957685, "epoch": 27.816039179675542, "frac_reward_zero_std": 0.15833333618938922, "grad_norm": 0.007049560546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4651403902.0, "reward": 8.188349628448487, "reward_std": 0.30748279541730883, "rewards/combined_reward/mean": 8.188349199295043, "rewards/combined_reward/std": 2.0601746678352355, "step": 45450 }, { "clip_ratio/high_max": 0.002237189729445769, "clip_ratio/high_mean": 0.0013444867974006533, "clip_ratio/low_mean": 0.0010333515419915784, "clip_ratio/low_min": 0.0003001631094472638, "clip_ratio/region_mean": 0.002377838340665524, "completions/clipped_ratio": 0.0, "completions/max_length": 587.6, "completions/max_terminated_length": 587.6, "completions/mean_length": 203.69531707763673, "completions/mean_terminated_length": 203.69531707763673, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02923913865815848, "epoch": 27.83440465258647, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.006072998046875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4654486252.0, "reward": 7.855830574035645, "reward_std": 0.31379175633192063, "rewards/combined_reward/mean": 7.855830287933349, "rewards/combined_reward/std": 2.4075783610343935, "step": 45480 }, { "clip_ratio/high_max": 0.0019874410356957623, "clip_ratio/high_mean": 0.0011868352059536847, "clip_ratio/low_mean": 0.0009259456539439271, "clip_ratio/low_min": 0.00029689941972416514, "clip_ratio/region_mean": 0.0021127808630505267, "completions/clipped_ratio": 0.0, "completions/max_length": 555.8, "completions/max_terminated_length": 555.8, "completions/mean_length": 201.21771392822265, "completions/mean_terminated_length": 201.21771392822265, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02661765863498052, "epoch": 27.852770125497397, "frac_reward_zero_std": 0.17916667349636556, "grad_norm": 0.023681640625, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4657559088.0, "reward": 7.8885101795196535, "reward_std": 0.3735572278499603, "rewards/combined_reward/mean": 7.888510131835938, "rewards/combined_reward/std": 2.25582594871521, "step": 45510 }, { "clip_ratio/high_max": 0.0029856492183171213, "clip_ratio/high_mean": 0.0016563486791710603, "clip_ratio/low_mean": 0.0011024351366359041, "clip_ratio/low_min": 0.0003507875611830968, "clip_ratio/region_mean": 0.002758783832541667, "completions/clipped_ratio": 0.0, "completions/max_length": 665.6, "completions/max_terminated_length": 665.6, "completions/mean_length": 200.34661865234375, "completions/mean_terminated_length": 200.34661865234375, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03351903008297086, "epoch": 27.871135598408326, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4660628579.0, "reward": 8.09228482246399, "reward_std": 0.37895721644163133, "rewards/combined_reward/mean": 8.092284631729125, "rewards/combined_reward/std": 2.27751544713974, "step": 45540 }, { "clip_ratio/high_max": 0.002183812054863665, "clip_ratio/high_mean": 0.0013218373578032091, "clip_ratio/low_mean": 0.001156434044848235, "clip_ratio/low_min": 0.0002696632422157563, "clip_ratio/region_mean": 0.002478271402651444, "completions/clipped_ratio": 0.0, "completions/max_length": 578.6, "completions/max_terminated_length": 578.6, "completions/mean_length": 203.4250045776367, "completions/mean_terminated_length": 203.4250045776367, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.02822719708395501, "epoch": 27.889501071319252, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4663709891.0, "reward": 7.776280927658081, "reward_std": 0.36845274865627287, "rewards/combined_reward/mean": 7.776280546188355, "rewards/combined_reward/std": 2.412202751636505, "step": 45570 }, { "clip_ratio/high_max": 0.0023604295347467996, "clip_ratio/high_mean": 0.0014756092215975514, "clip_ratio/low_mean": 0.0008879688364080115, "clip_ratio/low_min": 0.00020733433799856963, "clip_ratio/region_mean": 0.002363578057459866, "completions/clipped_ratio": 0.0, "completions/max_length": 600.2, "completions/max_terminated_length": 600.2, "completions/mean_length": 209.05703430175782, "completions/mean_terminated_length": 209.05703430175782, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.029916094057261942, "epoch": 27.90786654423018, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4666812830.0, "reward": 7.862382984161377, "reward_std": 0.38199578076601026, "rewards/combined_reward/mean": 7.8623827457427975, "rewards/combined_reward/std": 2.428169846534729, "step": 45600 }, { "clip_ratio/high_max": 0.00240600822774771, "clip_ratio/high_mean": 0.0013793452616179516, "clip_ratio/low_mean": 0.0010284327084567242, "clip_ratio/low_min": 0.00032742005399389505, "clip_ratio/region_mean": 0.0024077779618285906, "completions/clipped_ratio": 0.0, "completions/max_length": 586.6, "completions/max_terminated_length": 586.6, "completions/mean_length": 195.33255767822266, "completions/mean_terminated_length": 195.33255767822266, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.030366150274251897, "epoch": 27.926232017141107, "frac_reward_zero_std": 0.2041666701436043, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4669863067.0, "reward": 7.932655954360962, "reward_std": 0.3477974206209183, "rewards/combined_reward/mean": 7.9326558113098145, "rewards/combined_reward/std": 2.2943528175354, "step": 45630 }, { "clip_ratio/high_max": 0.0029324022432168323, "clip_ratio/high_mean": 0.0016912094829118966, "clip_ratio/low_mean": 0.000874290099091013, "clip_ratio/low_min": 0.00024526089036953637, "clip_ratio/region_mean": 0.002565499566359601, "completions/clipped_ratio": 0.0, "completions/max_length": 772.3, "completions/max_terminated_length": 772.3, "completions/mean_length": 203.3830795288086, "completions/mean_terminated_length": 203.3830795288086, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.032444927965601285, "epoch": 27.944597490052036, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4672944218.0, "reward": 7.805823850631714, "reward_std": 0.3885283231735229, "rewards/combined_reward/mean": 7.80582365989685, "rewards/combined_reward/std": 2.6569344878196715, "step": 45660 }, { "clip_ratio/high_max": 0.002679748979911286, "clip_ratio/high_mean": 0.0016706592636182905, "clip_ratio/low_mean": 0.000990034184117879, "clip_ratio/low_min": 0.00031259902583163543, "clip_ratio/region_mean": 0.002660693425665765, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 198.73568267822264, "completions/mean_terminated_length": 198.73568267822264, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.03270749460595349, "epoch": 27.962962962962962, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.01953125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4676007523.0, "reward": 8.096115779876708, "reward_std": 0.38004388511180875, "rewards/combined_reward/mean": 8.096115207672119, "rewards/combined_reward/std": 2.1747186303138735, "step": 45690 }, { "clip_ratio/high_max": 0.0025615465453787084, "clip_ratio/high_mean": 0.0014726891124155373, "clip_ratio/low_mean": 0.001003184991471547, "clip_ratio/low_min": 0.00031873529442236757, "clip_ratio/region_mean": 0.0024758740910328925, "completions/clipped_ratio": 0.0, "completions/max_length": 535.7, "completions/max_terminated_length": 535.7, "completions/mean_length": 203.32526550292968, "completions/mean_terminated_length": 203.32526550292968, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.02826456796998779, "epoch": 27.98132843587389, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4679088452.0, "reward": 8.000124454498291, "reward_std": 0.3671425670385361, "rewards/combined_reward/mean": 8.000124263763428, "rewards/combined_reward/std": 2.2390916466712953, "step": 45720 }, { "clip_ratio/high_max": 0.002245563268661499, "clip_ratio/high_mean": 0.0013237693437986309, "clip_ratio/low_mean": 0.0009870150403609539, "clip_ratio/low_min": 0.0002834887639134346, "clip_ratio/region_mean": 0.0023107843876156646, "completions/clipped_ratio": 0.0, "completions/max_length": 746.3, "completions/max_terminated_length": 746.3, "completions/mean_length": 193.4127655029297, "completions/mean_terminated_length": 193.4127655029297, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.030074980451414983, "epoch": 27.999693908784817, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 4682131317.0, "reward": 8.09056658744812, "reward_std": 0.2775152623653412, "rewards/combined_reward/mean": 8.090566205978394, "rewards/combined_reward/std": 2.142941927909851, "step": 45750 }, { "clip_ratio/high_max": 0.0019510850676343287, "clip_ratio/high_mean": 0.001129557736221922, "clip_ratio/low_mean": 0.0007776253359980281, "clip_ratio/low_min": 0.000237743229402024, "clip_ratio/region_mean": 0.0019071830704934516, "completions/clipped_ratio": 0.0, "completions/max_length": 666.7, "completions/max_terminated_length": 666.7, "completions/mean_length": 199.28724365234376, "completions/mean_terminated_length": 199.28724365234376, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.030351644445021274, "epoch": 28.017753290480563, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4685196740.0, "reward": 7.962167835235595, "reward_std": 0.28946504592895506, "rewards/combined_reward/mean": 7.962167501449585, "rewards/combined_reward/std": 2.258705532550812, "step": 45780 }, { "clip_ratio/high_max": 0.0016006931070781623, "clip_ratio/high_mean": 0.000965386172417008, "clip_ratio/low_mean": 0.0008617979863629443, "clip_ratio/low_min": 0.00018566552195503999, "clip_ratio/region_mean": 0.0018271841710278144, "completions/clipped_ratio": 0.0, "completions/max_length": 600.3, "completions/max_terminated_length": 600.3, "completions/mean_length": 201.45937957763672, "completions/mean_terminated_length": 201.45937957763672, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028234896482899785, "epoch": 28.036118763391492, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4688270504.0, "reward": 8.00237741470337, "reward_std": 0.29358205646276475, "rewards/combined_reward/mean": 8.002377080917359, "rewards/combined_reward/std": 2.2794207513332365, "step": 45810 }, { "clip_ratio/high_max": 0.002242686900717672, "clip_ratio/high_mean": 0.001179438985370022, "clip_ratio/low_mean": 0.0008287355747597757, "clip_ratio/low_min": 0.00019202093074757916, "clip_ratio/region_mean": 0.0020081745421824354, "completions/clipped_ratio": 0.0, "completions/max_length": 537.2, "completions/max_terminated_length": 537.2, "completions/mean_length": 194.60521392822267, "completions/mean_terminated_length": 194.60521392822267, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03061594950656096, "epoch": 28.054484236302418, "frac_reward_zero_std": 0.1416666716337204, "grad_norm": 0.0111083984375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4691317948.0, "reward": 8.007682514190673, "reward_std": 0.34666041284799576, "rewards/combined_reward/mean": 8.0076819896698, "rewards/combined_reward/std": 2.359084463119507, "step": 45840 }, { "clip_ratio/high_max": 0.002073608541589541, "clip_ratio/high_mean": 0.0010973272711756485, "clip_ratio/low_mean": 0.0008281073006704294, "clip_ratio/low_min": 0.00021925286588763508, "clip_ratio/region_mean": 0.001925434565055184, "completions/clipped_ratio": 0.0, "completions/max_length": 592.8, "completions/max_terminated_length": 592.8, "completions/mean_length": 185.49844207763672, "completions/mean_terminated_length": 185.49844207763672, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.029729194232883553, "epoch": 28.072849709213347, "frac_reward_zero_std": 0.2250000059604645, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4694330422.0, "reward": 8.2594482421875, "reward_std": 0.24564500600099565, "rewards/combined_reward/mean": 8.259447765350341, "rewards/combined_reward/std": 2.0999215841293335, "step": 45870 }, { "clip_ratio/high_max": 0.0018518026884218367, "clip_ratio/high_mean": 0.001057833314310604, "clip_ratio/low_mean": 0.0006533327739210411, "clip_ratio/low_min": 0.00016471214087990422, "clip_ratio/region_mean": 0.0017111660791366981, "completions/clipped_ratio": 0.0, "completions/max_length": 543.2, "completions/max_terminated_length": 543.2, "completions/mean_length": 193.43880615234374, "completions/mean_terminated_length": 193.43880615234374, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.025410490001862247, "epoch": 28.091215182124273, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.002593994140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4697373387.0, "reward": 8.235014390945434, "reward_std": 0.2768424741923809, "rewards/combined_reward/mean": 8.235014057159423, "rewards/combined_reward/std": 2.099648153781891, "step": 45900 }, { "clip_ratio/high_max": 0.0019422978926741052, "clip_ratio/high_mean": 0.0012568444926728262, "clip_ratio/low_mean": 0.0007214866437910435, "clip_ratio/low_min": 0.00022827286544876795, "clip_ratio/region_mean": 0.0019783311407081785, "completions/clipped_ratio": 0.0, "completions/max_length": 623.2, "completions/max_terminated_length": 623.2, "completions/mean_length": 202.3593780517578, "completions/mean_terminated_length": 202.3593780517578, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02857542977047463, "epoch": 28.1095806550352, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4700450607.0, "reward": 8.155512428283691, "reward_std": 0.33271016776561735, "rewards/combined_reward/mean": 8.155512046813964, "rewards/combined_reward/std": 2.204909086227417, "step": 45930 }, { "clip_ratio/high_max": 0.0017999328372146314, "clip_ratio/high_mean": 0.0011295533001733323, "clip_ratio/low_mean": 0.0007910921684015193, "clip_ratio/low_min": 0.000233972665349332, "clip_ratio/region_mean": 0.0019206454744562507, "completions/clipped_ratio": 0.0, "completions/max_length": 660.2, "completions/max_terminated_length": 660.2, "completions/mean_length": 206.0265670776367, "completions/mean_terminated_length": 206.0265670776367, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.02948258480367561, "epoch": 28.127946127946128, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4703541909.0, "reward": 8.072261142730714, "reward_std": 0.3272047497332096, "rewards/combined_reward/mean": 8.07226095199585, "rewards/combined_reward/std": 2.40809006690979, "step": 45960 }, { "clip_ratio/high_max": 0.0020859158738555077, "clip_ratio/high_mean": 0.0011578066133855222, "clip_ratio/low_mean": 0.0006626684890458516, "clip_ratio/low_min": 0.00018642500047765981, "clip_ratio/region_mean": 0.0018204750978232673, "completions/clipped_ratio": 0.0, "completions/max_length": 614.6, "completions/max_terminated_length": 614.6, "completions/mean_length": 198.96146392822266, "completions/mean_terminated_length": 198.96146392822266, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.028981912803525727, "epoch": 28.146311600857054, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.005645751953125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4706606081.0, "reward": 7.992230367660523, "reward_std": 0.304235702008009, "rewards/combined_reward/mean": 7.9922301292419435, "rewards/combined_reward/std": 2.2813392877578735, "step": 45990 }, { "clip_ratio/high_max": 0.0016007268328394274, "clip_ratio/high_mean": 0.0010420590446301504, "clip_ratio/low_mean": 0.0007556881044365582, "clip_ratio/low_min": 0.0002869309013476595, "clip_ratio/region_mean": 0.001797747150703799, "completions/clipped_ratio": 0.0, "completions/max_length": 540.4285714285714, "completions/max_terminated_length": 540.4285714285714, "completions/mean_length": 195.66444178989954, "completions/mean_terminated_length": 195.66444178989954, "completions/min_length": 37.285714285714285, "completions/min_terminated_length": 37.285714285714285, "entropy": 0.028306361217983068, "epoch": 28.16406489133762, "frac_reward_zero_std": 0.18452381449086325, "grad_norm": 0.048583984375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4709701669.0, "reward": 8.28163937159947, "reward_std": 0.2928979354245322, "rewards/combined_reward/mean": 8.281639099121094, "rewards/combined_reward/std": 2.094034876142229, "step": 46020 }, { "clip_ratio/high_max": 0.0020415294920288337, "clip_ratio/high_mean": 0.0011693303877109429, "clip_ratio/low_mean": 0.000763602159228564, "clip_ratio/low_min": 0.0002181645172337691, "clip_ratio/region_mean": 0.0019329325334789852, "completions/clipped_ratio": 0.0, "completions/max_length": 655.4, "completions/max_terminated_length": 655.4, "completions/mean_length": 193.3578140258789, "completions/mean_terminated_length": 193.3578140258789, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.029230493353679776, "epoch": 28.182430364248546, "frac_reward_zero_std": 0.1916666693985462, "grad_norm": 0.0162353515625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4712744323.0, "reward": 8.067633152008057, "reward_std": 0.33800305873155595, "rewards/combined_reward/mean": 8.067633152008057, "rewards/combined_reward/std": 2.314678394794464, "step": 46050 }, { "clip_ratio/high_max": 0.0016802639232385748, "clip_ratio/high_mean": 0.0009304999421146931, "clip_ratio/low_mean": 0.0007316312484666317, "clip_ratio/low_min": 0.00019278353477905814, "clip_ratio/region_mean": 0.0016621311908238568, "completions/clipped_ratio": 0.0, "completions/max_length": 673.9, "completions/max_terminated_length": 673.9, "completions/mean_length": 187.23151702880858, "completions/mean_terminated_length": 187.23151702880858, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02891781454284986, "epoch": 28.200795837159472, "frac_reward_zero_std": 0.18333334028720855, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4715763452.0, "reward": 7.941171646118164, "reward_std": 0.3639385625720024, "rewards/combined_reward/mean": 7.941171646118164, "rewards/combined_reward/std": 2.303593873977661, "step": 46080 }, { "clip_ratio/high_max": 0.0018328929586762873, "clip_ratio/high_mean": 0.001091716906861014, "clip_ratio/low_mean": 0.0006993726560419115, "clip_ratio/low_min": 0.000161399380643464, "clip_ratio/region_mean": 0.0017910895655707767, "completions/clipped_ratio": 0.0, "completions/max_length": 548.8, "completions/max_terminated_length": 548.8, "completions/mean_length": 197.25469207763672, "completions/mean_terminated_length": 197.25469207763672, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.027038053593908748, "epoch": 28.2191613100704, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4718821070.0, "reward": 7.974589872360229, "reward_std": 0.3555558145046234, "rewards/combined_reward/mean": 7.974589538574219, "rewards/combined_reward/std": 2.4044609665870667, "step": 46110 }, { "clip_ratio/high_max": 0.0018232130707474426, "clip_ratio/high_mean": 0.001027369285535921, "clip_ratio/low_mean": 0.0006415080907269537, "clip_ratio/low_min": 0.00012829790333247122, "clip_ratio/region_mean": 0.001668877402456322, "completions/clipped_ratio": 0.0, "completions/max_length": 626.7, "completions/max_terminated_length": 626.7, "completions/mean_length": 203.53125457763673, "completions/mean_terminated_length": 203.53125457763673, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02732457605501016, "epoch": 28.237526782981327, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.04345703125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4721902790.0, "reward": 8.236363458633424, "reward_std": 0.34332557320594786, "rewards/combined_reward/mean": 8.236363172531128, "rewards/combined_reward/std": 2.0863138914108275, "step": 46140 }, { "clip_ratio/high_max": 0.0019266942030905436, "clip_ratio/high_mean": 0.0010868381112838202, "clip_ratio/low_mean": 0.0007090466092146623, "clip_ratio/low_min": 0.00022082158999789196, "clip_ratio/region_mean": 0.00179588471291936, "completions/clipped_ratio": 0.0, "completions/max_length": 560.5, "completions/max_terminated_length": 560.5, "completions/mean_length": 186.88203659057618, "completions/mean_terminated_length": 186.88203659057618, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.02844829383927087, "epoch": 28.255892255892256, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.00408935546875, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4724920577.0, "reward": 7.941918754577637, "reward_std": 0.4053514748811722, "rewards/combined_reward/mean": 7.941918563842774, "rewards/combined_reward/std": 2.45944539308548, "step": 46170 }, { "clip_ratio/high_max": 0.0022101349027555746, "clip_ratio/high_mean": 0.0012230739572260062, "clip_ratio/low_mean": 0.000780212214946611, "clip_ratio/low_min": 0.00028026395739289, "clip_ratio/region_mean": 0.002003286181328197, "completions/clipped_ratio": 0.0, "completions/max_length": 582.2, "completions/max_terminated_length": 582.2, "completions/mean_length": 190.01953887939453, "completions/mean_terminated_length": 190.01953887939453, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030786800337955354, "epoch": 28.274257728803182, "frac_reward_zero_std": 0.1791666742414236, "grad_norm": 0.0174560546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4727950412.0, "reward": 8.153856134414672, "reward_std": 0.3130591958761215, "rewards/combined_reward/mean": 8.153855895996093, "rewards/combined_reward/std": 2.27493417263031, "step": 46200 }, { "clip_ratio/high_max": 0.0021258382383772794, "clip_ratio/high_mean": 0.0011975314525140373, "clip_ratio/low_mean": 0.0007676247230241036, "clip_ratio/low_min": 0.00027096014043005804, "clip_ratio/region_mean": 0.001965156187361572, "completions/clipped_ratio": 0.0, "completions/max_length": 626.3, "completions/max_terminated_length": 626.3, "completions/mean_length": 196.32865295410156, "completions/mean_terminated_length": 196.32865295410156, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.030993170524016022, "epoch": 28.29262320171411, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.0037689208984375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4731004474.0, "reward": 7.880714368820191, "reward_std": 0.3518915385007858, "rewards/combined_reward/mean": 7.880714368820191, "rewards/combined_reward/std": 2.3910069465637207, "step": 46230 }, { "clip_ratio/high_max": 0.0020299396545548614, "clip_ratio/high_mean": 0.0011591857740616736, "clip_ratio/low_mean": 0.0007445033184922068, "clip_ratio/low_min": 0.000254094281262951, "clip_ratio/region_mean": 0.0019036891040741465, "completions/clipped_ratio": 0.0, "completions/max_length": 597.6, "completions/max_terminated_length": 597.6, "completions/mean_length": 201.49844207763672, "completions/mean_terminated_length": 201.49844207763672, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030282534596820673, "epoch": 28.310988674625037, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4734078388.0, "reward": 7.978855180740356, "reward_std": 0.34946520030498507, "rewards/combined_reward/mean": 7.978854942321777, "rewards/combined_reward/std": 2.4096924304962157, "step": 46260 }, { "clip_ratio/high_max": 0.0019044050607287014, "clip_ratio/high_mean": 0.0011585611578387518, "clip_ratio/low_mean": 0.0007446261322911595, "clip_ratio/low_min": 0.0001830622253085797, "clip_ratio/region_mean": 0.0019031872929190286, "completions/clipped_ratio": 0.0, "completions/max_length": 584.7, "completions/max_terminated_length": 584.7, "completions/mean_length": 195.32396392822267, "completions/mean_terminated_length": 195.32396392822267, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.030268387589603662, "epoch": 28.329354147535966, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4737128592.0, "reward": 7.906911611557007, "reward_std": 0.41478689312934874, "rewards/combined_reward/mean": 7.906911182403564, "rewards/combined_reward/std": 2.3652308344841004, "step": 46290 }, { "clip_ratio/high_max": 0.0019362308414808164, "clip_ratio/high_mean": 0.0011249302415914523, "clip_ratio/low_mean": 0.0007060932261083508, "clip_ratio/low_min": 0.00018243285764280397, "clip_ratio/region_mean": 0.0018310234697613244, "completions/clipped_ratio": 0.0, "completions/max_length": 568.3, "completions/max_terminated_length": 568.3, "completions/mean_length": 208.38776397705078, "completions/mean_terminated_length": 208.38776397705078, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.02833473078596095, "epoch": 28.34771962044689, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.007568359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4740228961.0, "reward": 8.22486457824707, "reward_std": 0.3600175827741623, "rewards/combined_reward/mean": 8.224864435195922, "rewards/combined_reward/std": 2.092189133167267, "step": 46320 }, { "clip_ratio/high_max": 0.0018131499743806974, "clip_ratio/high_mean": 0.0010180002685956425, "clip_ratio/low_mean": 0.000851167590493181, "clip_ratio/low_min": 0.000243213363137329, "clip_ratio/region_mean": 0.001869167875459728, "completions/clipped_ratio": 0.0, "completions/max_length": 571.5, "completions/max_terminated_length": 571.5, "completions/mean_length": 195.44375457763672, "completions/mean_terminated_length": 195.44375457763672, "completions/min_length": 48.2, "completions/min_terminated_length": 48.2, "entropy": 0.02784599900866548, "epoch": 28.36608509335782, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4743279625.0, "reward": 8.186230230331422, "reward_std": 0.34486826658248904, "rewards/combined_reward/mean": 8.186230134963989, "rewards/combined_reward/std": 2.0534903407096863, "step": 46350 }, { "clip_ratio/high_max": 0.002210141772714754, "clip_ratio/high_mean": 0.0012323233184967344, "clip_ratio/low_mean": 0.000940138494358204, "clip_ratio/low_min": 0.00026945542000855006, "clip_ratio/region_mean": 0.002172461820494694, "completions/clipped_ratio": 0.0, "completions/max_length": 558.1, "completions/max_terminated_length": 558.1, "completions/mean_length": 200.66354522705078, "completions/mean_terminated_length": 200.66354522705078, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03128218244140347, "epoch": 28.384450566268747, "frac_reward_zero_std": 0.1875000037252903, "grad_norm": 0.014404296875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4746350333.0, "reward": 8.006286239624023, "reward_std": 0.3202576547861099, "rewards/combined_reward/mean": 8.006286001205444, "rewards/combined_reward/std": 2.344105803966522, "step": 46380 }, { "clip_ratio/high_max": 0.0020127647342936448, "clip_ratio/high_mean": 0.0011496589456025201, "clip_ratio/low_mean": 0.0008231503613690923, "clip_ratio/low_min": 0.00021126311670135087, "clip_ratio/region_mean": 0.0019728093184918786, "completions/clipped_ratio": 0.0, "completions/max_length": 555.2, "completions/max_terminated_length": 555.2, "completions/mean_length": 206.25104522705078, "completions/mean_terminated_length": 206.25104522705078, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02873260749814411, "epoch": 28.402816039179676, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.0057373046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4749442497.0, "reward": 8.14018955230713, "reward_std": 0.2833716630935669, "rewards/combined_reward/mean": 8.140189409255981, "rewards/combined_reward/std": 2.2059683799743652, "step": 46410 }, { "clip_ratio/high_max": 0.001793058963570123, "clip_ratio/high_mean": 0.0010682545454377154, "clip_ratio/low_mean": 0.0009247090844534492, "clip_ratio/low_min": 0.00027374135243007913, "clip_ratio/region_mean": 0.0019929636201898877, "completions/clipped_ratio": 0.0, "completions/max_length": 650.3, "completions/max_terminated_length": 650.3, "completions/mean_length": 207.2776092529297, "completions/mean_terminated_length": 207.2776092529297, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.031579622874657313, "epoch": 28.4211815120906, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.005584716796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4752538603.0, "reward": 8.175335836410522, "reward_std": 0.33734296560287474, "rewards/combined_reward/mean": 8.17533540725708, "rewards/combined_reward/std": 2.3174701929092407, "step": 46440 }, { "clip_ratio/high_max": 0.002177676365924223, "clip_ratio/high_mean": 0.001301820241981962, "clip_ratio/low_mean": 0.0008099200125191904, "clip_ratio/low_min": 0.00021391537375166082, "clip_ratio/region_mean": 0.0021117402696593976, "completions/clipped_ratio": 0.0, "completions/max_length": 602.2, "completions/max_terminated_length": 602.2, "completions/mean_length": 200.87709045410156, "completions/mean_terminated_length": 200.87709045410156, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.03300572345033288, "epoch": 28.43954698500153, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.040771484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4755610131.0, "reward": 8.045248460769653, "reward_std": 0.3701592177152634, "rewards/combined_reward/mean": 8.045248365402221, "rewards/combined_reward/std": 2.4452335357666017, "step": 46470 }, { "clip_ratio/high_max": 0.0016910684964386747, "clip_ratio/high_mean": 0.0009294492161037245, "clip_ratio/low_mean": 0.0005989394667873664, "clip_ratio/low_min": 0.00016576457176900778, "clip_ratio/region_mean": 0.001528388673129181, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 204.52292022705078, "completions/mean_terminated_length": 204.52292022705078, "completions/min_length": 44.2, "completions/min_terminated_length": 44.2, "entropy": 0.024336407845839857, "epoch": 28.457912457912457, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4758695659.0, "reward": 8.136902332305908, "reward_std": 0.32442822158336637, "rewards/combined_reward/mean": 8.136902379989625, "rewards/combined_reward/std": 2.147833001613617, "step": 46500 }, { "clip_ratio/high_max": 0.0020155505980559004, "clip_ratio/high_mean": 0.0012003622551371032, "clip_ratio/low_mean": 0.0007517007133477212, "clip_ratio/low_min": 0.0001820867883604175, "clip_ratio/region_mean": 0.0019520629892213037, "completions/clipped_ratio": 0.0, "completions/max_length": 639.7, "completions/max_terminated_length": 639.7, "completions/mean_length": 203.82031860351563, "completions/mean_terminated_length": 203.82031860351563, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.030121135773758093, "epoch": 28.476277930823386, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.0036468505859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4761778489.0, "reward": 7.806448316574096, "reward_std": 0.37623003125190735, "rewards/combined_reward/mean": 7.806447887420655, "rewards/combined_reward/std": 2.515554630756378, "step": 46530 }, { "clip_ratio/high_max": 0.0019423067380557768, "clip_ratio/high_mean": 0.001053972637237166, "clip_ratio/low_mean": 0.0007214739654349008, "clip_ratio/low_min": 0.00019838000007439406, "clip_ratio/region_mean": 0.0017754466006105454, "completions/clipped_ratio": 0.0, "completions/max_length": 543.6, "completions/max_terminated_length": 543.6, "completions/mean_length": 199.9955795288086, "completions/mean_terminated_length": 199.9955795288086, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028730901842936873, "epoch": 28.49464340373431, "frac_reward_zero_std": 0.17500000223517417, "grad_norm": 0.0054931640625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4764846632.0, "reward": 8.239236640930176, "reward_std": 0.29858761131763456, "rewards/combined_reward/mean": 8.23923625946045, "rewards/combined_reward/std": 2.058031690120697, "step": 46560 }, { "clip_ratio/high_max": 0.001733578133280389, "clip_ratio/high_mean": 0.0009269725237876021, "clip_ratio/low_mean": 0.0007031110943595801, "clip_ratio/low_min": 0.00017504302862410745, "clip_ratio/region_mean": 0.0016300836277271932, "completions/clipped_ratio": 0.0, "completions/max_length": 607.3, "completions/max_terminated_length": 607.3, "completions/mean_length": 202.71536712646486, "completions/mean_terminated_length": 202.71536712646486, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.02616167776286602, "epoch": 28.51300887664524, "frac_reward_zero_std": 0.17916667461395264, "grad_norm": 0.0096435546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4767925219.0, "reward": 8.09630708694458, "reward_std": 0.3403359979391098, "rewards/combined_reward/mean": 8.096306896209716, "rewards/combined_reward/std": 2.198963928222656, "step": 46590 }, { "clip_ratio/high_max": 0.0021127147738297936, "clip_ratio/high_mean": 0.0011706299366778695, "clip_ratio/low_mean": 0.0007243853547455122, "clip_ratio/low_min": 0.00017852131422841923, "clip_ratio/region_mean": 0.0018950152889980624, "completions/clipped_ratio": 0.0, "completions/max_length": 554.2, "completions/max_terminated_length": 554.2, "completions/mean_length": 197.25833892822266, "completions/mean_terminated_length": 197.25833892822266, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.02898101177997887, "epoch": 28.531374349556167, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 1.3515625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 4770982851.0, "reward": 7.964386463165283, "reward_std": 0.42275744676589966, "rewards/combined_reward/mean": 7.964386224746704, "rewards/combined_reward/std": 2.4403775095939637, "step": 46620 }, { "clip_ratio/high_max": 0.0018705084638592476, "clip_ratio/high_mean": 0.001104535062040668, "clip_ratio/low_mean": 0.0008156746517973564, "clip_ratio/low_min": 0.00022728545761007505, "clip_ratio/region_mean": 0.0019202097260858863, "completions/clipped_ratio": 0.0, "completions/max_length": 540.3, "completions/max_terminated_length": 540.3, "completions/mean_length": 192.4705780029297, "completions/mean_terminated_length": 192.4705780029297, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02535902298986912, "epoch": 28.549739822467096, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.028076171875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4774022098.0, "reward": 7.93141360282898, "reward_std": 0.32564807683229446, "rewards/combined_reward/mean": 7.931413221359253, "rewards/combined_reward/std": 2.3485357999801635, "step": 46650 }, { "clip_ratio/high_max": 0.0022617392368071403, "clip_ratio/high_mean": 0.0013142827452005197, "clip_ratio/low_mean": 0.0008895050044884556, "clip_ratio/low_min": 0.0003413423985572687, "clip_ratio/region_mean": 0.002203787747809353, "completions/clipped_ratio": 0.0, "completions/max_length": 612.2, "completions/max_terminated_length": 612.2, "completions/mean_length": 217.87187805175782, "completions/mean_terminated_length": 217.87187805175782, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.031601130170747636, "epoch": 28.56810529537802, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4777158886.0, "reward": 8.009414577484131, "reward_std": 0.38335442543029785, "rewards/combined_reward/mean": 8.009414577484131, "rewards/combined_reward/std": 2.22950222492218, "step": 46680 }, { "clip_ratio/high_max": 0.0022621689102379607, "clip_ratio/high_mean": 0.0013717232407846798, "clip_ratio/low_mean": 0.0009009977230258907, "clip_ratio/low_min": 0.0002743534741360539, "clip_ratio/region_mean": 0.0022727209642956344, "completions/clipped_ratio": 0.0, "completions/max_length": 627.6, "completions/max_terminated_length": 627.6, "completions/mean_length": 211.44427642822265, "completions/mean_terminated_length": 211.44427642822265, "completions/min_length": 44.1, "completions/min_terminated_length": 44.1, "entropy": 0.034278395523627596, "epoch": 28.58647076828895, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0107421875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4780270992.0, "reward": 8.223885011672973, "reward_std": 0.39293487519025805, "rewards/combined_reward/mean": 8.223884725570679, "rewards/combined_reward/std": 2.1759204149246214, "step": 46710 }, { "clip_ratio/high_max": 0.0021482682117493824, "clip_ratio/high_mean": 0.0012895276940980693, "clip_ratio/low_mean": 0.0008072051849012496, "clip_ratio/low_min": 0.0002497516875640334, "clip_ratio/region_mean": 0.002096732900827192, "completions/clipped_ratio": 0.0, "completions/max_length": 589.4, "completions/max_terminated_length": 589.4, "completions/mean_length": 209.79974517822265, "completions/mean_terminated_length": 209.79974517822265, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03140675324636201, "epoch": 28.604836241199877, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4783376783.0, "reward": 8.086403512954712, "reward_std": 0.32949815690517426, "rewards/combined_reward/mean": 8.086403131484985, "rewards/combined_reward/std": 2.196392822265625, "step": 46740 }, { "clip_ratio/high_max": 0.0020376366849329014, "clip_ratio/high_mean": 0.0011584274485358036, "clip_ratio/low_mean": 0.0007675290925059623, "clip_ratio/low_min": 0.00021233037114143372, "clip_ratio/region_mean": 0.001925956531583021, "completions/clipped_ratio": 0.0, "completions/max_length": 549.1, "completions/max_terminated_length": 549.1, "completions/mean_length": 190.0916717529297, "completions/mean_terminated_length": 190.0916717529297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029755592710959414, "epoch": 28.623201714110806, "frac_reward_zero_std": 0.21250000596046448, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4786406895.0, "reward": 7.7722008228302, "reward_std": 0.33766874223947524, "rewards/combined_reward/mean": 7.77220048904419, "rewards/combined_reward/std": 2.5162264585494993, "step": 46770 }, { "clip_ratio/high_max": 0.0021231473656371236, "clip_ratio/high_mean": 0.0012748259049355207, "clip_ratio/low_mean": 0.0008798730896766453, "clip_ratio/low_min": 0.0002702364470072401, "clip_ratio/region_mean": 0.0021546990004329323, "completions/clipped_ratio": 0.0, "completions/max_length": 518.9, "completions/max_terminated_length": 518.9, "completions/mean_length": 198.32031860351563, "completions/mean_terminated_length": 198.32031860351563, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.02868959077168256, "epoch": 28.64156718702173, "frac_reward_zero_std": 0.13750000447034835, "grad_norm": 0.01397705078125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4789468605.0, "reward": 8.143038654327393, "reward_std": 0.4046799838542938, "rewards/combined_reward/mean": 8.143038415908814, "rewards/combined_reward/std": 2.072869873046875, "step": 46800 }, { "clip_ratio/high_max": 0.0018241638506879098, "clip_ratio/high_mean": 0.0010728739342691065, "clip_ratio/low_mean": 0.0007130881555591865, "clip_ratio/low_min": 0.00019667343488739182, "clip_ratio/region_mean": 0.0017859621065629957, "completions/clipped_ratio": 0.0, "completions/max_length": 676.8, "completions/max_terminated_length": 676.8, "completions/mean_length": 205.35755615234376, "completions/mean_terminated_length": 205.35755615234376, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029687890775191288, "epoch": 28.65993265993266, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0439453125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4792557338.0, "reward": 8.131954574584961, "reward_std": 0.33520656228065493, "rewards/combined_reward/mean": 8.131954383850097, "rewards/combined_reward/std": 2.343129301071167, "step": 46830 }, { "clip_ratio/high_max": 0.002033350320562022, "clip_ratio/high_mean": 0.001054206816115766, "clip_ratio/low_mean": 0.0007400142511566325, "clip_ratio/low_min": 0.0002485396408398325, "clip_ratio/region_mean": 0.0017942210676361963, "completions/clipped_ratio": 0.0, "completions/max_length": 622.2, "completions/max_terminated_length": 622.2, "completions/mean_length": 190.15573425292968, "completions/mean_terminated_length": 190.15573425292968, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02779704096416632, "epoch": 28.678298132843587, "frac_reward_zero_std": 0.22083333879709244, "grad_norm": 0.007354736328125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4795587696.0, "reward": 7.961623954772949, "reward_std": 0.29924906194210055, "rewards/combined_reward/mean": 7.961623668670654, "rewards/combined_reward/std": 2.286229372024536, "step": 46860 }, { "clip_ratio/high_max": 0.0020580528192416145, "clip_ratio/high_mean": 0.0010745829982624854, "clip_ratio/low_mean": 0.0007467013401765143, "clip_ratio/low_min": 0.00018336378003975067, "clip_ratio/region_mean": 0.001821284338075202, "completions/clipped_ratio": 0.0, "completions/max_length": 607.7, "completions/max_terminated_length": 607.7, "completions/mean_length": 187.15443420410156, "completions/mean_terminated_length": 187.15443420410156, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.02684890989524623, "epoch": 28.696663605754516, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.005157470703125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4798606529.0, "reward": 8.116336584091187, "reward_std": 0.28947755694389343, "rewards/combined_reward/mean": 8.116336441040039, "rewards/combined_reward/std": 2.3050053834915163, "step": 46890 }, { "clip_ratio/high_max": 0.00207009971345542, "clip_ratio/high_mean": 0.0012220898774103262, "clip_ratio/low_mean": 0.0008917282979382435, "clip_ratio/low_min": 0.00022016743702503542, "clip_ratio/region_mean": 0.0021138181822607295, "completions/clipped_ratio": 0.0, "completions/max_length": 633.9, "completions/max_terminated_length": 633.9, "completions/mean_length": 212.1502670288086, "completions/mean_terminated_length": 212.1502670288086, "completions/min_length": 46.4, "completions/min_terminated_length": 46.4, "entropy": 0.03082582486482958, "epoch": 28.71502907866544, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0255126953125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4801721346.0, "reward": 7.938853025436401, "reward_std": 0.38527142703533174, "rewards/combined_reward/mean": 7.938852691650391, "rewards/combined_reward/std": 2.478282558917999, "step": 46920 }, { "clip_ratio/high_max": 0.0016971093027677853, "clip_ratio/high_mean": 0.0009735810455216172, "clip_ratio/low_mean": 0.0007027783782784051, "clip_ratio/low_min": 0.00020255291747162119, "clip_ratio/region_mean": 0.0016763594282868629, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 198.89037017822267, "completions/mean_terminated_length": 198.89037017822267, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027597619600904486, "epoch": 28.73339455157637, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4804785245.0, "reward": 7.942247152328491, "reward_std": 0.2972835436463356, "rewards/combined_reward/mean": 7.942246818542481, "rewards/combined_reward/std": 2.394572651386261, "step": 46950 }, { "clip_ratio/high_max": 0.0016314404291430642, "clip_ratio/high_mean": 0.0009913757882410816, "clip_ratio/low_mean": 0.0007270806437494078, "clip_ratio/low_min": 0.00022617856278278243, "clip_ratio/region_mean": 0.0017184564232593403, "completions/clipped_ratio": 0.0, "completions/max_length": 550.7, "completions/max_terminated_length": 550.7, "completions/mean_length": 204.07318115234375, "completions/mean_terminated_length": 204.07318115234375, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027815485171352825, "epoch": 28.751760024487297, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4807869046.0, "reward": 8.25270643234253, "reward_std": 0.2848301030695438, "rewards/combined_reward/mean": 8.252705955505371, "rewards/combined_reward/std": 2.0302637338638307, "step": 46980 }, { "clip_ratio/high_max": 0.0020240779500454663, "clip_ratio/high_mean": 0.0011177068395530417, "clip_ratio/low_mean": 0.0007999989626114257, "clip_ratio/low_min": 0.00016405178466811777, "clip_ratio/region_mean": 0.0019177057954948395, "completions/clipped_ratio": 0.0, "completions/max_length": 596.8, "completions/max_terminated_length": 596.8, "completions/mean_length": 192.2487045288086, "completions/mean_terminated_length": 192.2487045288086, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029305544417972364, "epoch": 28.770125497398226, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4810907441.0, "reward": 8.205499696731568, "reward_std": 0.29396957233548165, "rewards/combined_reward/mean": 8.205499696731568, "rewards/combined_reward/std": 2.117144060134888, "step": 47010 }, { "clip_ratio/high_max": 0.0018608266259737623, "clip_ratio/high_mean": 0.0010425677624880335, "clip_ratio/low_mean": 0.0007978374897599376, "clip_ratio/low_min": 0.00018227770172719223, "clip_ratio/region_mean": 0.0018404052452145455, "completions/clipped_ratio": 0.0, "completions/max_length": 553.4, "completions/max_terminated_length": 553.4, "completions/mean_length": 198.8783920288086, "completions/mean_terminated_length": 198.8783920288086, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.028936774314691622, "epoch": 28.78849097030915, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0213623046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4813971294.0, "reward": 8.198911952972413, "reward_std": 0.2911056637763977, "rewards/combined_reward/mean": 8.19891185760498, "rewards/combined_reward/std": 2.149349129199982, "step": 47040 }, { "clip_ratio/high_max": 0.0019927443863707595, "clip_ratio/high_mean": 0.0011043340139925324, "clip_ratio/low_mean": 0.0007546725334880951, "clip_ratio/low_min": 0.0001837428584015773, "clip_ratio/region_mean": 0.001859006542751255, "completions/clipped_ratio": 0.0, "completions/max_length": 591.5, "completions/max_terminated_length": 591.5, "completions/mean_length": 207.34271545410155, "completions/mean_terminated_length": 207.34271545410155, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.0283178490276138, "epoch": 28.80685644322008, "frac_reward_zero_std": 0.15000000558793544, "grad_norm": 0.0260009765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4817067650.0, "reward": 8.313862562179565, "reward_std": 0.3175933748483658, "rewards/combined_reward/mean": 8.313862228393555, "rewards/combined_reward/std": 2.0427319288253782, "step": 47070 }, { "clip_ratio/high_max": 0.0019566639205246855, "clip_ratio/high_mean": 0.0010720712189746942, "clip_ratio/low_mean": 0.00087412380950506, "clip_ratio/low_min": 0.0002804457260936033, "clip_ratio/region_mean": 0.0019461950165956903, "completions/clipped_ratio": 0.0, "completions/max_length": 538.4, "completions/max_terminated_length": 538.4, "completions/mean_length": 201.86901397705077, "completions/mean_terminated_length": 201.86901397705077, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.028511263150721788, "epoch": 28.825221916131007, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.005096435546875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4820142987.0, "reward": 8.131070280075074, "reward_std": 0.267092115432024, "rewards/combined_reward/mean": 8.131070375442505, "rewards/combined_reward/std": 2.1157554984092712, "step": 47100 }, { "clip_ratio/high_max": 0.001893312793011622, "clip_ratio/high_mean": 0.0011352569209217714, "clip_ratio/low_mean": 0.000889865983845084, "clip_ratio/low_min": 0.00030817166286093806, "clip_ratio/region_mean": 0.002025122872631376, "completions/clipped_ratio": 0.0, "completions/max_length": 635.6, "completions/max_terminated_length": 635.6, "completions/mean_length": 190.6744842529297, "completions/mean_terminated_length": 190.6744842529297, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.029965573720013104, "epoch": 28.843587389041936, "frac_reward_zero_std": 0.23750001043081284, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4823175337.0, "reward": 7.936294460296631, "reward_std": 0.3733726918697357, "rewards/combined_reward/mean": 7.936294317245483, "rewards/combined_reward/std": 2.297128736972809, "step": 47130 }, { "clip_ratio/high_max": 0.0019396509446475345, "clip_ratio/high_mean": 0.00111318485190471, "clip_ratio/low_mean": 0.0006753765964579846, "clip_ratio/low_min": 0.0001662214330281131, "clip_ratio/region_mean": 0.0017885614584277694, "completions/clipped_ratio": 0.0, "completions/max_length": 579.4, "completions/max_terminated_length": 579.4, "completions/mean_length": 205.18646392822265, "completions/mean_terminated_length": 205.18646392822265, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.028825715153167644, "epoch": 28.86195286195286, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0272216796875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4826263413.0, "reward": 8.263995695114136, "reward_std": 0.3554891675710678, "rewards/combined_reward/mean": 8.26399531364441, "rewards/combined_reward/std": 2.0641920804977416, "step": 47160 }, { "clip_ratio/high_max": 0.0020167291678565864, "clip_ratio/high_mean": 0.0011183570418021797, "clip_ratio/low_mean": 0.0008649880367253596, "clip_ratio/low_min": 0.0002672484935222504, "clip_ratio/region_mean": 0.0019833450933219865, "completions/clipped_ratio": 0.0, "completions/max_length": 606.4, "completions/max_terminated_length": 606.4, "completions/mean_length": 195.79063110351564, "completions/mean_terminated_length": 195.79063110351564, "completions/min_length": 39.2, "completions/min_terminated_length": 39.2, "entropy": 0.030742462538182736, "epoch": 28.88031833486379, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.005767822265625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4829315409.0, "reward": 8.211728143692017, "reward_std": 0.37317543625831606, "rewards/combined_reward/mean": 8.211728048324584, "rewards/combined_reward/std": 2.137668716907501, "step": 47190 }, { "clip_ratio/high_max": 0.002166677520532782, "clip_ratio/high_mean": 0.001148066187791604, "clip_ratio/low_mean": 0.0007635486943399882, "clip_ratio/low_min": 0.00016626712240395137, "clip_ratio/region_mean": 0.0019116148629109374, "completions/clipped_ratio": 0.0, "completions/max_length": 594.3, "completions/max_terminated_length": 594.3, "completions/mean_length": 203.16172485351564, "completions/mean_terminated_length": 203.16172485351564, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.030246490178008876, "epoch": 28.898683807774717, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.029296875, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 4832395710.0, "reward": 7.984961318969726, "reward_std": 0.4358807623386383, "rewards/combined_reward/mean": 7.984961128234863, "rewards/combined_reward/std": 2.3361886620521544, "step": 47220 }, { "clip_ratio/high_max": 0.0022737438572221435, "clip_ratio/high_mean": 0.0013971836645699416, "clip_ratio/low_mean": 0.0007924627747949368, "clip_ratio/low_min": 0.0002557854177818323, "clip_ratio/region_mean": 0.002189646443972985, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 200.34219360351562, "completions/mean_terminated_length": 200.34219360351562, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03266480692351858, "epoch": 28.917049280685646, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4835465184.0, "reward": 8.133777332305907, "reward_std": 0.3132824406027794, "rewards/combined_reward/mean": 8.133777236938476, "rewards/combined_reward/std": 2.3177839040756227, "step": 47250 }, { "clip_ratio/high_max": 0.0018414584919810295, "clip_ratio/high_mean": 0.0010609955120647403, "clip_ratio/low_mean": 0.0007448075445912158, "clip_ratio/low_min": 0.00020548750690068117, "clip_ratio/region_mean": 0.001805803060415201, "completions/clipped_ratio": 0.0, "completions/max_length": 552.2, "completions/max_terminated_length": 552.2, "completions/mean_length": 183.6174530029297, "completions/mean_terminated_length": 183.6174530029297, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.02569366372190416, "epoch": 28.93541475359657, "frac_reward_zero_std": 0.18750000819563867, "grad_norm": 0.00439453125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4838470435.0, "reward": 7.8228835105896, "reward_std": 0.3244502291083336, "rewards/combined_reward/mean": 7.822883415222168, "rewards/combined_reward/std": 2.4593198895454407, "step": 47280 }, { "clip_ratio/high_max": 0.0018865945351232464, "clip_ratio/high_mean": 0.0011535623819023991, "clip_ratio/low_mean": 0.0008166628738157063, "clip_ratio/low_min": 0.00026111462987804166, "clip_ratio/region_mean": 0.001970225251473797, "completions/clipped_ratio": 0.0, "completions/max_length": 609.5, "completions/max_terminated_length": 609.5, "completions/mean_length": 212.46537017822266, "completions/mean_terminated_length": 212.46537017822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.0302948211474965, "epoch": 28.9537802265075, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.023193359375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4841586462.0, "reward": 8.047394752502441, "reward_std": 0.30626251697540285, "rewards/combined_reward/mean": 8.047394561767579, "rewards/combined_reward/std": 2.178382956981659, "step": 47310 }, { "clip_ratio/high_max": 0.0020628074150105626, "clip_ratio/high_mean": 0.0012237480921612586, "clip_ratio/low_mean": 0.0007593353623330283, "clip_ratio/low_min": 0.0002558170097472612, "clip_ratio/region_mean": 0.0019830834692887342, "completions/clipped_ratio": 0.0, "completions/max_length": 540.9, "completions/max_terminated_length": 540.9, "completions/mean_length": 198.7135482788086, "completions/mean_terminated_length": 198.7135482788086, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029621286721279224, "epoch": 28.972145699418427, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4844649682.0, "reward": 7.917226839065552, "reward_std": 0.3345481395721436, "rewards/combined_reward/mean": 7.91722674369812, "rewards/combined_reward/std": 2.438541758060455, "step": 47340 }, { "clip_ratio/high_max": 0.0018955316632248771, "clip_ratio/high_mean": 0.0010554599751534017, "clip_ratio/low_mean": 0.0008414534597856497, "clip_ratio/low_min": 0.00024434951192233714, "clip_ratio/region_mean": 0.0018969134466412167, "completions/clipped_ratio": 0.0, "completions/max_length": 635.7, "completions/max_terminated_length": 635.7, "completions/mean_length": 202.80209045410157, "completions/mean_terminated_length": 202.80209045410157, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029395615697527924, "epoch": 28.990511172329356, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4847728602.0, "reward": 7.873862934112549, "reward_std": 0.3197010524570942, "rewards/combined_reward/mean": 7.87386269569397, "rewards/combined_reward/std": 2.5117199301719664, "step": 47370 }, { "clip_ratio/high_max": 0.0023481747675610735, "clip_ratio/high_mean": 0.001279621594012944, "clip_ratio/low_mean": 0.0007461955840985642, "clip_ratio/low_min": 0.00018334871307816968, "clip_ratio/region_mean": 0.0020258171831211837, "completions/clipped_ratio": 0.0, "completions/max_length": 669.7, "completions/max_terminated_length": 669.7, "completions/mean_length": 203.2502670288086, "completions/mean_terminated_length": 203.2502670288086, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02795184139528724, "epoch": 29.009182736455465, "frac_reward_zero_std": 0.17083334028720856, "grad_norm": 0.02685546875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4850809243.0, "reward": 8.2085045337677, "reward_std": 0.3171134442090988, "rewards/combined_reward/mean": 8.208504056930542, "rewards/combined_reward/std": 1.9574198007583619, "step": 47400 }, { "clip_ratio/high_max": 0.0022197149189499515, "clip_ratio/high_mean": 0.0013115850480971858, "clip_ratio/low_mean": 0.0010170290822619184, "clip_ratio/low_min": 0.0002715396617228786, "clip_ratio/region_mean": 0.00232861412417454, "completions/clipped_ratio": 0.0, "completions/max_length": 513.3, "completions/max_terminated_length": 513.3, "completions/mean_length": 201.4432357788086, "completions/mean_terminated_length": 201.4432357788086, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.02811691085807979, "epoch": 29.02754820936639, "frac_reward_zero_std": 0.20000000707805157, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 4853882945.0, "reward": 8.019813585281373, "reward_std": 0.2423749178647995, "rewards/combined_reward/mean": 8.019813537597656, "rewards/combined_reward/std": 2.2815683364868162, "step": 47430 }, { "clip_ratio/high_max": 0.0026991778121252235, "clip_ratio/high_mean": 0.0014706771148970195, "clip_ratio/low_mean": 0.0010769411104168587, "clip_ratio/low_min": 0.0003683190145238768, "clip_ratio/region_mean": 0.0025476182282242613, "completions/clipped_ratio": 0.0, "completions/max_length": 592.7, "completions/max_terminated_length": 592.7, "completions/mean_length": 209.4174545288086, "completions/mean_terminated_length": 209.4174545288086, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.031357558971891804, "epoch": 29.04591368227732, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4856987268.0, "reward": 7.907603454589844, "reward_std": 0.3750214733183384, "rewards/combined_reward/mean": 7.907603359222412, "rewards/combined_reward/std": 2.5155653595924377, "step": 47460 }, { "clip_ratio/high_max": 0.0027761023006557177, "clip_ratio/high_mean": 0.001667541261728426, "clip_ratio/low_mean": 0.0010546452265164893, "clip_ratio/low_min": 0.0002940230037590178, "clip_ratio/region_mean": 0.0027221864943082134, "completions/clipped_ratio": 0.0, "completions/max_length": 748.2, "completions/max_terminated_length": 748.2, "completions/mean_length": 211.93125762939454, "completions/mean_terminated_length": 211.93125762939454, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.030079982553919155, "epoch": 29.064279155188245, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.022705078125, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 4860101244.0, "reward": 8.061839294433593, "reward_std": 0.383264946937561, "rewards/combined_reward/mean": 8.061838865280151, "rewards/combined_reward/std": 2.305621635913849, "step": 47490 }, { "clip_ratio/high_max": 0.0024823382525937633, "clip_ratio/high_mean": 0.0014968424115068047, "clip_ratio/low_mean": 0.0009188121711000955, "clip_ratio/low_min": 0.00026206871407339347, "clip_ratio/region_mean": 0.0024156545841833577, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 213.01458587646485, "completions/mean_terminated_length": 213.01458587646485, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.028616329034169514, "epoch": 29.082644628099175, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 4863219380.0, "reward": 7.98083643913269, "reward_std": 0.30809353590011596, "rewards/combined_reward/mean": 7.9808361530303955, "rewards/combined_reward/std": 2.3570491075515747, "step": 47520 }, { "clip_ratio/high_max": 0.002024549138635242, "clip_ratio/high_mean": 0.0011172403012096766, "clip_ratio/low_mean": 0.0008633332763565704, "clip_ratio/low_min": 0.0002297341292432975, "clip_ratio/region_mean": 0.001980573592542593, "completions/clipped_ratio": 0.0, "completions/max_length": 521.2, "completions/max_terminated_length": 521.2, "completions/mean_length": 192.73125457763672, "completions/mean_terminated_length": 192.73125457763672, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02521398583582292, "epoch": 29.1010101010101, "frac_reward_zero_std": 0.1916666742414236, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4866259628.0, "reward": 8.158610439300537, "reward_std": 0.32021884322166444, "rewards/combined_reward/mean": 8.158609771728516, "rewards/combined_reward/std": 2.268090319633484, "step": 47550 }, { "clip_ratio/high_max": 0.0022219740858417936, "clip_ratio/high_mean": 0.0012410314634810978, "clip_ratio/low_mean": 0.0009969638667826075, "clip_ratio/low_min": 0.0003118682482939524, "clip_ratio/region_mean": 0.0022379953385097905, "completions/clipped_ratio": 0.0, "completions/max_length": 615.1, "completions/max_terminated_length": 615.1, "completions/mean_length": 205.0260452270508, "completions/mean_terminated_length": 205.0260452270508, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027268327384566266, "epoch": 29.11937557392103, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.00579833984375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4869347088.0, "reward": 8.333412504196167, "reward_std": 0.29801507741212846, "rewards/combined_reward/mean": 8.33341188430786, "rewards/combined_reward/std": 2.0555501222610473, "step": 47580 }, { "clip_ratio/high_max": 0.0021953273962329453, "clip_ratio/high_mean": 0.0012299681657168549, "clip_ratio/low_mean": 0.0009023422346217558, "clip_ratio/low_min": 0.00021554423631945003, "clip_ratio/region_mean": 0.0021323104165882492, "completions/clipped_ratio": 0.0, "completions/max_length": 598.1, "completions/max_terminated_length": 598.1, "completions/mean_length": 212.4088592529297, "completions/mean_terminated_length": 212.4088592529297, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.027421699309100708, "epoch": 29.137741046831955, "frac_reward_zero_std": 0.15416667237877846, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 4872462898.0, "reward": 8.157081460952758, "reward_std": 0.4043621629476547, "rewards/combined_reward/mean": 8.157081174850465, "rewards/combined_reward/std": 2.1254355907440186, "step": 47610 }, { "clip_ratio/high_max": 0.002754603847764277, "clip_ratio/high_mean": 0.0016029033292094634, "clip_ratio/low_mean": 0.0011081600131243854, "clip_ratio/low_min": 0.00031102589588651123, "clip_ratio/region_mean": 0.0027110633777435093, "completions/clipped_ratio": 0.0, "completions/max_length": 705.4, "completions/max_terminated_length": 705.4, "completions/mean_length": 194.1890670776367, "completions/mean_terminated_length": 194.1890670776367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03326044975159069, "epoch": 29.156106519742885, "frac_reward_zero_std": 0.20000000558793546, "grad_norm": 0.028564453125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4875508744.0, "reward": 8.120974397659301, "reward_std": 0.33381485641002656, "rewards/combined_reward/mean": 8.120974111557008, "rewards/combined_reward/std": 2.1071378350257874, "step": 47640 }, { "clip_ratio/high_max": 0.0027130563510581853, "clip_ratio/high_mean": 0.0015913175067301685, "clip_ratio/low_mean": 0.0011695444166737918, "clip_ratio/low_min": 0.0003398849832592532, "clip_ratio/region_mean": 0.0027608619294672582, "completions/clipped_ratio": 0.0, "completions/max_length": 576.1, "completions/max_terminated_length": 576.1, "completions/mean_length": 197.9914108276367, "completions/mean_terminated_length": 197.9914108276367, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.033491688501089814, "epoch": 29.17447199265381, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 4878569191.0, "reward": 7.701187658309936, "reward_std": 0.317451485991478, "rewards/combined_reward/mean": 7.701187515258789, "rewards/combined_reward/std": 2.608933997154236, "step": 47670 }, { "clip_ratio/high_max": 0.001987683949118946, "clip_ratio/high_mean": 0.0011668585729542732, "clip_ratio/low_mean": 0.0008968907989886551, "clip_ratio/low_min": 0.00019701038011893008, "clip_ratio/region_mean": 0.0020637493473865715, "completions/clipped_ratio": 0.0, "completions/max_length": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 195.46042175292968, "completions/mean_terminated_length": 195.46042175292968, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026301988827375073, "epoch": 29.19283746556474, "frac_reward_zero_std": 0.2083333395421505, "grad_norm": 0.026611328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4881619919.0, "reward": 7.995115900039673, "reward_std": 0.2732994332909584, "rewards/combined_reward/mean": 7.99511570930481, "rewards/combined_reward/std": 2.587096607685089, "step": 47700 }, { "clip_ratio/high_max": 0.002289643594122026, "clip_ratio/high_mean": 0.0012649601838347734, "clip_ratio/low_mean": 0.0007419406241145529, "clip_ratio/low_min": 0.00014473415673516382, "clip_ratio/region_mean": 0.0020069008045538795, "completions/clipped_ratio": 0.0, "completions/max_length": 574.4, "completions/max_terminated_length": 574.4, "completions/mean_length": 199.97344207763672, "completions/mean_terminated_length": 199.97344207763672, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.026392410229891538, "epoch": 29.211202938475665, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4884687977.0, "reward": 8.165435552597046, "reward_std": 0.3197905957698822, "rewards/combined_reward/mean": 8.16543526649475, "rewards/combined_reward/std": 2.1003663897514344, "step": 47730 }, { "clip_ratio/high_max": 0.0024695108048035765, "clip_ratio/high_mean": 0.0013808805898103552, "clip_ratio/low_mean": 0.000992336945516096, "clip_ratio/low_min": 0.00023152139183366672, "clip_ratio/region_mean": 0.0023732175497571005, "completions/clipped_ratio": 0.0, "completions/max_length": 588.3, "completions/max_terminated_length": 588.3, "completions/mean_length": 199.15521392822265, "completions/mean_terminated_length": 199.15521392822265, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.02887639553907017, "epoch": 29.229568411386595, "frac_reward_zero_std": 0.16666666865348817, "grad_norm": 0.03125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4887752893.0, "reward": 7.942142200469971, "reward_std": 0.26802645027637484, "rewards/combined_reward/mean": 7.942141819000244, "rewards/combined_reward/std": 2.397514271736145, "step": 47760 }, { "clip_ratio/high_max": 0.0025392338119369622, "clip_ratio/high_mean": 0.001495110965333879, "clip_ratio/low_mean": 0.0009537667136707266, "clip_ratio/low_min": 0.0002975807811405199, "clip_ratio/region_mean": 0.002448877681066127, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 202.76380767822266, "completions/mean_terminated_length": 202.76380767822266, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.029471334473540386, "epoch": 29.24793388429752, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01116943359375, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 4890831666.0, "reward": 8.166628551483154, "reward_std": 0.27715086936950684, "rewards/combined_reward/mean": 8.166628217697143, "rewards/combined_reward/std": 2.1825002431869507, "step": 47790 }, { "clip_ratio/high_max": 0.0023652556524514996, "clip_ratio/high_mean": 0.00148706007845855, "clip_ratio/low_mean": 0.0009975934502411594, "clip_ratio/low_min": 0.0002708569079307684, "clip_ratio/region_mean": 0.002484653540523141, "completions/clipped_ratio": 0.0, "completions/max_length": 588.9, "completions/max_terminated_length": 588.9, "completions/mean_length": 205.78359985351562, "completions/mean_terminated_length": 205.78359985351562, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02969098094229897, "epoch": 29.26629935720845, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4893922035.0, "reward": 8.28985595703125, "reward_std": 0.3643276885151863, "rewards/combined_reward/mean": 8.289855813980102, "rewards/combined_reward/std": 1.9646640419960022, "step": 47820 }, { "clip_ratio/high_max": 0.001946222662809305, "clip_ratio/high_mean": 0.0010852026435410759, "clip_ratio/low_mean": 0.0008252644524570012, "clip_ratio/low_min": 0.00024048461442968498, "clip_ratio/region_mean": 0.0019104671058206197, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 198.59323272705078, "completions/mean_terminated_length": 198.59323272705078, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02432664871836702, "epoch": 29.284664830119375, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4896984793.0, "reward": 8.218257713317872, "reward_std": 0.30047989040613177, "rewards/combined_reward/mean": 8.21825737953186, "rewards/combined_reward/std": 2.0180269360542296, "step": 47850 }, { "clip_ratio/high_max": 0.0021423160811536945, "clip_ratio/high_mean": 0.0011994130762711089, "clip_ratio/low_mean": 0.000912063422583742, "clip_ratio/low_min": 0.000135030496797602, "clip_ratio/region_mean": 0.0021114765001887764, "completions/clipped_ratio": 0.0, "completions/max_length": 590.4, "completions/max_terminated_length": 590.4, "completions/mean_length": 188.4674545288086, "completions/mean_terminated_length": 188.4674545288086, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.02547415429726243, "epoch": 29.303030303030305, "frac_reward_zero_std": 0.22083334028720855, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 4900008668.0, "reward": 8.317538166046143, "reward_std": 0.3654015712440014, "rewards/combined_reward/mean": 8.317537784576416, "rewards/combined_reward/std": 1.9999544024467468, "step": 47880 }, { "clip_ratio/high_max": 0.0031189954626218725, "clip_ratio/high_mean": 0.0016391012681803356, "clip_ratio/low_mean": 0.0009044634596648393, "clip_ratio/low_min": 0.0002593040090384117, "clip_ratio/region_mean": 0.0025435647345148026, "completions/clipped_ratio": 0.0, "completions/max_length": 636.3, "completions/max_terminated_length": 636.3, "completions/mean_length": 209.44271545410157, "completions/mean_terminated_length": 209.44271545410157, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.033143060010237, "epoch": 29.32139577594123, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4903113088.0, "reward": 8.017725849151612, "reward_std": 0.3448526903986931, "rewards/combined_reward/mean": 8.017725610733033, "rewards/combined_reward/std": 2.3062468647956846, "step": 47910 }, { "clip_ratio/high_max": 0.0023033434801618567, "clip_ratio/high_mean": 0.0013605664559387758, "clip_ratio/low_mean": 0.0008971743257764805, "clip_ratio/low_min": 0.0002526001442068567, "clip_ratio/region_mean": 0.0022577407624339687, "completions/clipped_ratio": 0.0, "completions/max_length": 609.3, "completions/max_terminated_length": 609.3, "completions/mean_length": 205.6526092529297, "completions/mean_terminated_length": 205.6526092529297, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.027616236296792825, "epoch": 29.33976124885216, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.0089111328125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 4906202954.0, "reward": 7.906085586547851, "reward_std": 0.3387486129999161, "rewards/combined_reward/mean": 7.906085348129272, "rewards/combined_reward/std": 2.6218507289886475, "step": 47940 }, { "clip_ratio/high_max": 0.002289692221772081, "clip_ratio/high_mean": 0.0014354778184497263, "clip_ratio/low_mean": 0.001029722961660203, "clip_ratio/low_min": 0.0003380248027193981, "clip_ratio/region_mean": 0.002465200759858514, "completions/clipped_ratio": 0.0, "completions/max_length": 638.3, "completions/max_terminated_length": 638.3, "completions/mean_length": 208.83099517822265, "completions/mean_terminated_length": 208.83099517822265, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "entropy": 0.029609441990032792, "epoch": 29.358126721763085, "frac_reward_zero_std": 0.17500000186264514, "grad_norm": 0.0322265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4909305025.0, "reward": 8.022542047500611, "reward_std": 0.39689556658267977, "rewards/combined_reward/mean": 8.022541618347168, "rewards/combined_reward/std": 2.2394107580184937, "step": 47970 }, { "clip_ratio/high_max": 0.0024102043251332363, "clip_ratio/high_mean": 0.0013583336908292646, "clip_ratio/low_mean": 0.0009343744451446885, "clip_ratio/low_min": 0.0002766177550559708, "clip_ratio/region_mean": 0.0022927081496163736, "completions/clipped_ratio": 0.0, "completions/max_length": 606.4, "completions/max_terminated_length": 606.4, "completions/mean_length": 200.17682647705078, "completions/mean_terminated_length": 200.17682647705078, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028171110525727273, "epoch": 29.376492194674015, "frac_reward_zero_std": 0.17916666939854622, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 4912373864.0, "reward": 8.348797464370728, "reward_std": 0.30971541404724123, "rewards/combined_reward/mean": 8.348797273635864, "rewards/combined_reward/std": 1.8391430735588075, "step": 48000 }, { "clip_ratio/high_max": 0.002058064113225555, "clip_ratio/high_mean": 0.0011990198519924889, "clip_ratio/low_mean": 0.0006804622154353031, "clip_ratio/low_min": 0.0002072536236179682, "clip_ratio/region_mean": 0.0018794820537247384, "completions/clipped_ratio": 0.0, "completions/max_length": 593.8, "completions/max_terminated_length": 593.8, "completions/mean_length": 208.5416687011719, "completions/mean_terminated_length": 208.5416687011719, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.02944637704640627, "epoch": 29.394245485154578, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4915474824.0, "reward": 8.031286144256592, "reward_std": 0.3128997251391411, "rewards/combined_reward/mean": 8.031285953521728, "rewards/combined_reward/std": 2.3019837379455566, "step": 48030 }, { "clip_ratio/high_max": 0.0018045992226689122, "clip_ratio/high_mean": 0.0010768678951232384, "clip_ratio/low_mean": 0.0007336541831440021, "clip_ratio/low_min": 0.00017711580342923602, "clip_ratio/region_mean": 0.0018105220816020543, "completions/clipped_ratio": 0.0, "completions/max_length": 648.5, "completions/max_terminated_length": 648.5, "completions/mean_length": 209.6658905029297, "completions/mean_terminated_length": 209.6658905029297, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.02893808113100628, "epoch": 29.412610958065503, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4918580101.0, "reward": 8.211305713653564, "reward_std": 0.3462166041135788, "rewards/combined_reward/mean": 8.211305475234985, "rewards/combined_reward/std": 2.173534667491913, "step": 48060 }, { "clip_ratio/high_max": 0.0022846324022490687, "clip_ratio/high_mean": 0.0012208630883833393, "clip_ratio/low_mean": 0.0007645473722125947, "clip_ratio/low_min": 0.00026193581264427244, "clip_ratio/region_mean": 0.0019854104495607315, "completions/clipped_ratio": 0.0, "completions/max_length": 698.7, "completions/max_terminated_length": 698.7, "completions/mean_length": 201.2526107788086, "completions/mean_terminated_length": 201.2526107788086, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.03398554896314939, "epoch": 29.430976430976433, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 4921653071.0, "reward": 8.075520133972168, "reward_std": 0.32548277378082274, "rewards/combined_reward/mean": 8.075519847869874, "rewards/combined_reward/std": 2.423589849472046, "step": 48090 }, { "clip_ratio/high_max": 0.0016002244752598926, "clip_ratio/high_mean": 0.0009343627746905744, "clip_ratio/low_mean": 0.0006565421627480343, "clip_ratio/low_min": 0.00015731810514504712, "clip_ratio/region_mean": 0.0015909049305870818, "completions/clipped_ratio": 0.0, "completions/max_length": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 199.61641387939454, "completions/mean_terminated_length": 199.61641387939454, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.025789498630911113, "epoch": 29.44934190388736, "frac_reward_zero_std": 0.12083333469927311, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4924719758.0, "reward": 8.04531774520874, "reward_std": 0.3639805570244789, "rewards/combined_reward/mean": 8.04531741142273, "rewards/combined_reward/std": 2.328804135322571, "step": 48120 }, { "clip_ratio/high_max": 0.0019076023861998692, "clip_ratio/high_mean": 0.0011073511329110868, "clip_ratio/low_mean": 0.0007654939855153013, "clip_ratio/low_min": 0.00021544706178246997, "clip_ratio/region_mean": 0.0018728450920510416, "completions/clipped_ratio": 0.0, "completions/max_length": 649.4, "completions/max_terminated_length": 649.4, "completions/mean_length": 216.7369873046875, "completions/mean_terminated_length": 216.7369873046875, "completions/min_length": 47.2, "completions/min_terminated_length": 47.2, "entropy": 0.028828553576022386, "epoch": 29.467707376798288, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.01031494140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4927852188.0, "reward": 7.984402799606324, "reward_std": 0.3588805735111237, "rewards/combined_reward/mean": 7.984402465820312, "rewards/combined_reward/std": 2.3410610437393187, "step": 48150 }, { "clip_ratio/high_max": 0.001769316779973451, "clip_ratio/high_mean": 0.0010673457058146597, "clip_ratio/low_mean": 0.0007424366149886434, "clip_ratio/low_min": 0.00012479303914005868, "clip_ratio/region_mean": 0.0018097823282005264, "completions/clipped_ratio": 0.0, "completions/max_length": 588.9, "completions/max_terminated_length": 588.9, "completions/mean_length": 194.59766235351563, "completions/mean_terminated_length": 194.59766235351563, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.026861932982380193, "epoch": 29.486072849709213, "frac_reward_zero_std": 0.20833333916962146, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 4930899603.0, "reward": 8.063007688522339, "reward_std": 0.31149185001850127, "rewards/combined_reward/mean": 8.063007402420045, "rewards/combined_reward/std": 2.280620515346527, "step": 48180 }, { "clip_ratio/high_max": 0.0020387468153785448, "clip_ratio/high_mean": 0.001094461907390117, "clip_ratio/low_mean": 0.0007594557417178293, "clip_ratio/low_min": 0.00022494441849024344, "clip_ratio/region_mean": 0.001853917644378574, "completions/clipped_ratio": 0.0, "completions/max_length": 561.4, "completions/max_terminated_length": 561.4, "completions/mean_length": 202.44193115234376, "completions/mean_terminated_length": 202.44193115234376, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02969913707735638, "epoch": 29.50443832262014, "frac_reward_zero_std": 0.1666666693985462, "grad_norm": 0.037109375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4933977140.0, "reward": 8.089887523651123, "reward_std": 0.3392515152692795, "rewards/combined_reward/mean": 8.089887380599976, "rewards/combined_reward/std": 2.1097679138183594, "step": 48210 }, { "clip_ratio/high_max": 0.001648281278903596, "clip_ratio/high_mean": 0.0009513178725076917, "clip_ratio/low_mean": 0.0006743977491472226, "clip_ratio/low_min": 0.00015525036287726833, "clip_ratio/region_mean": 0.0016257156215336483, "completions/clipped_ratio": 0.0, "completions/max_length": 596.7, "completions/max_terminated_length": 596.7, "completions/mean_length": 198.31224670410157, "completions/mean_terminated_length": 198.31224670410157, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.02521727865872284, "epoch": 29.52280379553107, "frac_reward_zero_std": 0.18333333618938924, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 4937038819.0, "reward": 8.11105637550354, "reward_std": 0.38370162695646287, "rewards/combined_reward/mean": 8.111055898666383, "rewards/combined_reward/std": 2.2182466268539427, "step": 48240 }, { "clip_ratio/high_max": 0.0019623619838966986, "clip_ratio/high_mean": 0.0010984349093632773, "clip_ratio/low_mean": 0.0008271821185189765, "clip_ratio/low_min": 0.0003082759370348261, "clip_ratio/region_mean": 0.0019256170387961901, "completions/clipped_ratio": 0.0, "completions/max_length": 583.1, "completions/max_terminated_length": 583.1, "completions/mean_length": 200.66771545410157, "completions/mean_terminated_length": 200.66771545410157, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02901581779588014, "epoch": 29.541169268441998, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.037841796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4940109543.0, "reward": 7.946041297912598, "reward_std": 0.3502121567726135, "rewards/combined_reward/mean": 7.946040964126587, "rewards/combined_reward/std": 2.4561590433120726, "step": 48270 }, { "clip_ratio/high_max": 0.0017372087992650147, "clip_ratio/high_mean": 0.0010722415708490493, "clip_ratio/low_mean": 0.0007490558307229852, "clip_ratio/low_min": 0.00020793300257840504, "clip_ratio/region_mean": 0.0018212974226723114, "completions/clipped_ratio": 0.0, "completions/max_length": 576.6, "completions/max_terminated_length": 576.6, "completions/mean_length": 201.83698120117188, "completions/mean_terminated_length": 201.83698120117188, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.026669557920346657, "epoch": 29.559534741352923, "frac_reward_zero_std": 0.20000000670552254, "grad_norm": 0.031982421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4943184757.0, "reward": 7.983572149276734, "reward_std": 0.322384187579155, "rewards/combined_reward/mean": 7.983572149276734, "rewards/combined_reward/std": 2.2378528833389284, "step": 48300 }, { "clip_ratio/high_max": 0.00223606015254821, "clip_ratio/high_mean": 0.0013571392812688524, "clip_ratio/low_mean": 0.0009000994511249397, "clip_ratio/low_min": 0.0002958089188420369, "clip_ratio/region_mean": 0.002257238720388462, "completions/clipped_ratio": 0.0, "completions/max_length": 626.4, "completions/max_terminated_length": 626.4, "completions/mean_length": 217.9054748535156, "completions/mean_terminated_length": 217.9054748535156, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.03722125492058694, "epoch": 29.57790021426385, "frac_reward_zero_std": 0.12083333544433117, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4946321674.0, "reward": 8.08314232826233, "reward_std": 0.3728511810302734, "rewards/combined_reward/mean": 8.083141994476318, "rewards/combined_reward/std": 2.2687251925468446, "step": 48330 }, { "clip_ratio/high_max": 0.0019538611736303817, "clip_ratio/high_mean": 0.0011503843816171866, "clip_ratio/low_mean": 0.000754606020442831, "clip_ratio/low_min": 0.0001647487370064482, "clip_ratio/region_mean": 0.0019049903964817835, "completions/clipped_ratio": 0.0, "completions/max_length": 746.4, "completions/max_terminated_length": 746.4, "completions/mean_length": 204.01744995117187, "completions/mean_terminated_length": 204.01744995117187, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027782243746332824, "epoch": 29.59626568717478, "frac_reward_zero_std": 0.16666667126119136, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0031, "num_tokens": 4949405261.0, "reward": 8.186825704574584, "reward_std": 0.3685866117477417, "rewards/combined_reward/mean": 8.18682541847229, "rewards/combined_reward/std": 2.1586262822151183, "step": 48360 }, { "clip_ratio/high_max": 0.002399836554832291, "clip_ratio/high_mean": 0.001407559220145534, "clip_ratio/low_mean": 0.0009467455677319473, "clip_ratio/low_min": 0.00035347053587126234, "clip_ratio/region_mean": 0.002354304762169098, "completions/clipped_ratio": 0.0, "completions/max_length": 607.5, "completions/max_terminated_length": 607.5, "completions/mean_length": 198.06354827880858, "completions/mean_terminated_length": 198.06354827880858, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.032464374074091515, "epoch": 29.614631160085704, "frac_reward_zero_std": 0.18750000484287738, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 4952465985.0, "reward": 7.980441761016846, "reward_std": 0.3295851916074753, "rewards/combined_reward/mean": 7.980441856384277, "rewards/combined_reward/std": 2.33502254486084, "step": 48390 }, { "clip_ratio/high_max": 0.0018367860539001413, "clip_ratio/high_mean": 0.0010497351913121141, "clip_ratio/low_mean": 0.000801051092154618, "clip_ratio/low_min": 0.00019542821052406605, "clip_ratio/region_mean": 0.0018507862880748386, "completions/clipped_ratio": 0.0, "completions/max_length": 538.1, "completions/max_terminated_length": 538.1, "completions/mean_length": 197.39349517822265, "completions/mean_terminated_length": 197.39349517822265, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.028438616489681105, "epoch": 29.632996632996633, "frac_reward_zero_std": 0.19583333656191826, "grad_norm": 0.0262451171875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4955524136.0, "reward": 7.885916328430175, "reward_std": 0.2940741077065468, "rewards/combined_reward/mean": 7.885916042327881, "rewards/combined_reward/std": 2.2754154205322266, "step": 48420 }, { "clip_ratio/high_max": 0.0019117995024619935, "clip_ratio/high_mean": 0.0010417950514238327, "clip_ratio/low_mean": 0.0007000002688679767, "clip_ratio/low_min": 0.00021295226009290975, "clip_ratio/region_mean": 0.0017417953289016925, "completions/clipped_ratio": 0.0, "completions/max_length": 563.5, "completions/max_terminated_length": 563.5, "completions/mean_length": 199.97266235351563, "completions/mean_terminated_length": 199.97266235351563, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02766965899306039, "epoch": 29.65136210590756, "frac_reward_zero_std": 0.15000000186264514, "grad_norm": 0.0050048828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 4958592191.0, "reward": 8.198213338851929, "reward_std": 0.33768985122442247, "rewards/combined_reward/mean": 8.198212957382202, "rewards/combined_reward/std": 2.147496449947357, "step": 48450 }, { "clip_ratio/high_max": 0.0015767491456548063, "clip_ratio/high_mean": 0.0009307090746005997, "clip_ratio/low_mean": 0.0007973131014296087, "clip_ratio/low_min": 0.0001751483629050199, "clip_ratio/region_mean": 0.0017280221945839003, "completions/clipped_ratio": 0.0, "completions/max_length": 616.1, "completions/max_terminated_length": 616.1, "completions/mean_length": 193.6408905029297, "completions/mean_terminated_length": 193.6408905029297, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02718464849361529, "epoch": 29.669727578818488, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.00958251953125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4961635932.0, "reward": 7.933099412918091, "reward_std": 0.30323600172996523, "rewards/combined_reward/mean": 7.933099317550659, "rewards/combined_reward/std": 2.451450192928314, "step": 48480 }, { "clip_ratio/high_max": 0.002095557955908589, "clip_ratio/high_mean": 0.0011353017271176213, "clip_ratio/low_mean": 0.0005848311886438751, "clip_ratio/low_min": 0.0001236160380358342, "clip_ratio/region_mean": 0.0017201329178836507, "completions/clipped_ratio": 0.0, "completions/max_length": 609.3, "completions/max_terminated_length": 609.3, "completions/mean_length": 190.4875061035156, "completions/mean_terminated_length": 190.4875061035156, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.025636917476852737, "epoch": 29.688093051729414, "frac_reward_zero_std": 0.2166666716337204, "grad_norm": 0.0059814453125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4964667564.0, "reward": 8.115714406967163, "reward_std": 0.29861471205949786, "rewards/combined_reward/mean": 8.115713930130005, "rewards/combined_reward/std": 2.2044796705245973, "step": 48510 }, { "clip_ratio/high_max": 0.002095428881875705, "clip_ratio/high_mean": 0.0012089958310146662, "clip_ratio/low_mean": 0.0008403120464208769, "clip_ratio/low_min": 0.00023427456108038314, "clip_ratio/region_mean": 0.0020493078719785746, "completions/clipped_ratio": 0.0, "completions/max_length": 672.6, "completions/max_terminated_length": 672.6, "completions/mean_length": 204.9237045288086, "completions/mean_terminated_length": 204.9237045288086, "completions/min_length": 41.1, "completions/min_terminated_length": 41.1, "entropy": 0.031618221880247196, "epoch": 29.706458524640343, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4967754631.0, "reward": 8.096235370635986, "reward_std": 0.3419667214155197, "rewards/combined_reward/mean": 8.096235370635986, "rewards/combined_reward/std": 2.3535291075706484, "step": 48540 }, { "clip_ratio/high_max": 0.0020516690958174875, "clip_ratio/high_mean": 0.001203631449849733, "clip_ratio/low_mean": 0.000712615043084952, "clip_ratio/low_min": 0.00015906181821871238, "clip_ratio/region_mean": 0.001916246495966334, "completions/clipped_ratio": 0.0, "completions/max_length": 637.5, "completions/max_terminated_length": 637.5, "completions/mean_length": 198.75157012939454, "completions/mean_terminated_length": 198.75157012939454, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.02950915270484984, "epoch": 29.72482399755127, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 4970817997.0, "reward": 7.822881031036377, "reward_std": 0.32572810649871825, "rewards/combined_reward/mean": 7.822880649566651, "rewards/combined_reward/std": 2.5395251154899596, "step": 48570 }, { "clip_ratio/high_max": 0.0016859859955729918, "clip_ratio/high_mean": 0.0008953733309075081, "clip_ratio/low_mean": 0.0007785252028649363, "clip_ratio/low_min": 0.0002319295505003538, "clip_ratio/region_mean": 0.0016738985267390186, "completions/clipped_ratio": 0.0, "completions/max_length": 538.5, "completions/max_terminated_length": 538.5, "completions/mean_length": 195.5276077270508, "completions/mean_terminated_length": 195.5276077270508, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.025141094640518228, "epoch": 29.743189470462198, "frac_reward_zero_std": 0.1916666742414236, "grad_norm": 0.00299072265625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 4973868983.0, "reward": 8.169502401351929, "reward_std": 0.2598863676190376, "rewards/combined_reward/mean": 8.16950216293335, "rewards/combined_reward/std": 2.0518265783786775, "step": 48600 }, { "clip_ratio/high_max": 0.002328149366561168, "clip_ratio/high_mean": 0.0012505833323909125, "clip_ratio/low_mean": 0.0008160118953431569, "clip_ratio/low_min": 0.0002063288139955451, "clip_ratio/region_mean": 0.0020665952096654413, "completions/clipped_ratio": 0.0, "completions/max_length": 601.2, "completions/max_terminated_length": 601.2, "completions/mean_length": 205.4330795288086, "completions/mean_terminated_length": 205.4330795288086, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03231493237738808, "epoch": 29.761554943373124, "frac_reward_zero_std": 0.18333334103226662, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 4976958006.0, "reward": 8.149277210235596, "reward_std": 0.32897928059101106, "rewards/combined_reward/mean": 8.149276685714721, "rewards/combined_reward/std": 2.1863759994506835, "step": 48630 }, { "clip_ratio/high_max": 0.0019900067374692297, "clip_ratio/high_mean": 0.001070829774592615, "clip_ratio/low_mean": 0.0008228705526562407, "clip_ratio/low_min": 0.00024131651734933258, "clip_ratio/region_mean": 0.001893700336222537, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 194.51745147705077, "completions/mean_terminated_length": 194.51745147705077, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02766507927638789, "epoch": 29.779920416284053, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.00543212890625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4980005113.0, "reward": 8.365055799484253, "reward_std": 0.32710598558187487, "rewards/combined_reward/mean": 8.365055418014526, "rewards/combined_reward/std": 1.9994537353515625, "step": 48660 }, { "clip_ratio/high_max": 0.001671496158329925, "clip_ratio/high_mean": 0.0009298565614396163, "clip_ratio/low_mean": 0.0007223672912611316, "clip_ratio/low_min": 0.00020931056424160487, "clip_ratio/region_mean": 0.001652223849305301, "completions/clipped_ratio": 0.0, "completions/max_length": 526.1, "completions/max_terminated_length": 526.1, "completions/mean_length": 202.56719207763672, "completions/mean_terminated_length": 202.56719207763672, "completions/min_length": 39.3, "completions/min_terminated_length": 39.3, "entropy": 0.02705496639634172, "epoch": 29.79828588919498, "frac_reward_zero_std": 0.1333333369344473, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4983083131.0, "reward": 8.098955965042114, "reward_std": 0.30423678904771806, "rewards/combined_reward/mean": 8.098955821990966, "rewards/combined_reward/std": 2.277357840538025, "step": 48690 }, { "clip_ratio/high_max": 0.001972568335865314, "clip_ratio/high_mean": 0.0010848552619184678, "clip_ratio/low_mean": 0.000833160757974838, "clip_ratio/low_min": 0.00023924236059732115, "clip_ratio/region_mean": 0.0019180160049775926, "completions/clipped_ratio": 0.0, "completions/max_length": 574.1, "completions/max_terminated_length": 574.1, "completions/mean_length": 203.89818267822267, "completions/mean_terminated_length": 203.89818267822267, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.027838812116533516, "epoch": 29.816651362105908, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 4986166260.0, "reward": 8.237188673019409, "reward_std": 0.29307708889245987, "rewards/combined_reward/mean": 8.237188673019409, "rewards/combined_reward/std": 2.11987122297287, "step": 48720 }, { "clip_ratio/high_max": 0.0019327260946738535, "clip_ratio/high_mean": 0.0011420621383877005, "clip_ratio/low_mean": 0.000872252738190582, "clip_ratio/low_min": 0.0003409456124548645, "clip_ratio/region_mean": 0.0020143148964659, "completions/clipped_ratio": 0.0, "completions/max_length": 548.4, "completions/max_terminated_length": 548.4, "completions/mean_length": 197.17031860351562, "completions/mean_terminated_length": 197.17031860351562, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.030417208070866765, "epoch": 29.835016835016834, "frac_reward_zero_std": 0.21250000447034836, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 4989223554.0, "reward": 7.896333312988281, "reward_std": 0.3317332223057747, "rewards/combined_reward/mean": 7.896332883834839, "rewards/combined_reward/std": 2.2851874828338623, "step": 48750 }, { "clip_ratio/high_max": 0.0018254899213692019, "clip_ratio/high_mean": 0.0010426387329061982, "clip_ratio/low_mean": 0.0007840235891004946, "clip_ratio/low_min": 0.0002225804865399065, "clip_ratio/region_mean": 0.001826662309031235, "completions/clipped_ratio": 0.0, "completions/max_length": 764.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 197.96589050292968, "completions/mean_terminated_length": 197.96589050292968, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.027260728180408478, "epoch": 29.853382307927763, "frac_reward_zero_std": 0.21666667237877846, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 4992283903.0, "reward": 8.273586511611938, "reward_std": 0.3289322406053543, "rewards/combined_reward/mean": 8.273586177825928, "rewards/combined_reward/std": 2.09629682302475, "step": 48780 }, { "clip_ratio/high_max": 0.0021276350894671244, "clip_ratio/high_mean": 0.0012215733918613599, "clip_ratio/low_mean": 0.0009006605262887509, "clip_ratio/low_min": 0.0003296174785646144, "clip_ratio/region_mean": 0.002122233935127345, "completions/clipped_ratio": 0.0, "completions/max_length": 582.4, "completions/max_terminated_length": 582.4, "completions/mean_length": 205.10364990234376, "completions/mean_terminated_length": 205.10364990234376, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.03421613983809948, "epoch": 29.87174778083869, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 4995371661.0, "reward": 8.241813325881958, "reward_std": 0.3345320463180542, "rewards/combined_reward/mean": 8.241812992095948, "rewards/combined_reward/std": 1.968183434009552, "step": 48810 }, { "clip_ratio/high_max": 0.0022818562826917815, "clip_ratio/high_mean": 0.0012783470214344562, "clip_ratio/low_mean": 0.0007428094003747294, "clip_ratio/low_min": 0.0001879765208286699, "clip_ratio/region_mean": 0.0020211564216879195, "completions/clipped_ratio": 0.0, "completions/max_length": 549.3, "completions/max_terminated_length": 549.3, "completions/mean_length": 192.6080795288086, "completions/mean_terminated_length": 192.6080795288086, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.029481911690284808, "epoch": 29.890113253749618, "frac_reward_zero_std": 0.19583334140479564, "grad_norm": 0.0211181640625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 4998411436.0, "reward": 8.02995057106018, "reward_std": 0.31683765202760694, "rewards/combined_reward/mean": 8.0299503326416, "rewards/combined_reward/std": 2.3721276044845583, "step": 48840 }, { "clip_ratio/high_max": 0.0020700113226970037, "clip_ratio/high_mean": 0.001216987663913945, "clip_ratio/low_mean": 0.000789406792682712, "clip_ratio/low_min": 0.00022497538181293447, "clip_ratio/region_mean": 0.0020063944481080397, "completions/clipped_ratio": 0.0, "completions/max_length": 576.2, "completions/max_terminated_length": 576.2, "completions/mean_length": 204.42760772705077, "completions/mean_terminated_length": 204.42760772705077, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02924185569087664, "epoch": 29.908478726660544, "frac_reward_zero_std": 0.18333333507180213, "grad_norm": 0.013671875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5001496598.0, "reward": 8.20658507347107, "reward_std": 0.3835479080677032, "rewards/combined_reward/mean": 8.206584596633911, "rewards/combined_reward/std": 2.0151838660240173, "step": 48870 }, { "clip_ratio/high_max": 0.0018232645884078617, "clip_ratio/high_mean": 0.0010772611898573815, "clip_ratio/low_mean": 0.0007459378403533871, "clip_ratio/low_min": 0.0001771746511318876, "clip_ratio/region_mean": 0.0018231990305745663, "completions/clipped_ratio": 0.0, "completions/max_length": 553.5, "completions/max_terminated_length": 553.5, "completions/mean_length": 195.6593795776367, "completions/mean_terminated_length": 195.6593795776367, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.029048738939066728, "epoch": 29.926844199571473, "frac_reward_zero_std": 0.22083334475755692, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5004548090.0, "reward": 7.9305041313171385, "reward_std": 0.2795905962586403, "rewards/combined_reward/mean": 7.930504035949707, "rewards/combined_reward/std": 2.5930598855018614, "step": 48900 }, { "clip_ratio/high_max": 0.0017937868765632932, "clip_ratio/high_mean": 0.0010241220086754765, "clip_ratio/low_mean": 0.0008134178555337712, "clip_ratio/low_min": 0.00025616844335066464, "clip_ratio/region_mean": 0.0018375398780335673, "completions/clipped_ratio": 0.0, "completions/max_length": 619.4, "completions/max_terminated_length": 619.4, "completions/mean_length": 190.19297485351564, "completions/mean_terminated_length": 190.19297485351564, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028205714002251626, "epoch": 29.9452096724824, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.00634765625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5007578591.0, "reward": 7.994154548645019, "reward_std": 0.313229139149189, "rewards/combined_reward/mean": 7.994154357910157, "rewards/combined_reward/std": 2.25188410282135, "step": 48930 }, { "clip_ratio/high_max": 0.002146577739525431, "clip_ratio/high_mean": 0.0012403407711341667, "clip_ratio/low_mean": 0.0008794117121093829, "clip_ratio/low_min": 0.0002695958777621854, "clip_ratio/region_mean": 0.0021197524891855817, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 212.2593795776367, "completions/mean_terminated_length": 212.2593795776367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029579329149176676, "epoch": 29.963575145393328, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 5010693827.0, "reward": 7.9540516376495365, "reward_std": 0.426294082403183, "rewards/combined_reward/mean": 7.954051113128662, "rewards/combined_reward/std": 2.1807103872299196, "step": 48960 }, { "clip_ratio/high_max": 0.0018125084701750894, "clip_ratio/high_mean": 0.0010339053785476911, "clip_ratio/low_mean": 0.0008448745130105333, "clip_ratio/low_min": 0.0002452683639906657, "clip_ratio/region_mean": 0.0018787798823420114, "completions/clipped_ratio": 0.0, "completions/max_length": 575.2, "completions/max_terminated_length": 575.2, "completions/mean_length": 194.68333892822267, "completions/mean_terminated_length": 194.68333892822267, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.030305550976966817, "epoch": 29.981940618304254, "frac_reward_zero_std": 0.21666667386889457, "grad_norm": 0.02685546875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5013741571.0, "reward": 7.913120174407959, "reward_std": 0.26628964245319364, "rewards/combined_reward/mean": 7.91311993598938, "rewards/combined_reward/std": 2.5086262226104736, "step": 48990 }, { "clip_ratio/high_max": 0.001677604206812904, "clip_ratio/high_mean": 0.000946199072460018, "clip_ratio/low_mean": 0.0006272586611348593, "clip_ratio/low_min": 0.00015832623132870944, "clip_ratio/region_mean": 0.001573457741646142, "completions/clipped_ratio": 0.0, "completions/max_length": 590.2727272727273, "completions/max_terminated_length": 590.2727272727273, "completions/mean_length": 202.70336497913706, "completions/mean_terminated_length": 202.70336497913706, "completions/min_length": 41.72727272727273, "completions/min_terminated_length": 41.72727272727273, "entropy": 0.025767641180179646, "epoch": 30.000612182430363, "frac_reward_zero_std": 0.1401515186510303, "grad_norm": 0.019287109375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5017127966.0, "reward": 8.054070646112615, "reward_std": 0.34067234938794916, "rewards/combined_reward/mean": 8.054070472717285, "rewards/combined_reward/std": 2.269826368852095, "step": 49020 }, { "clip_ratio/high_max": 0.002643317630281672, "clip_ratio/high_mean": 0.0014909927107510158, "clip_ratio/low_mean": 0.0010115791838567626, "clip_ratio/low_min": 0.00035022646795065764, "clip_ratio/region_mean": 0.0025025718874530866, "completions/clipped_ratio": 0.0, "completions/max_length": 694.7, "completions/max_terminated_length": 694.7, "completions/mean_length": 203.60547332763673, "completions/mean_terminated_length": 203.60547332763673, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02840071575095256, "epoch": 30.018977655341292, "frac_reward_zero_std": 0.2125000074505806, "grad_norm": 0.01544189453125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5020209971.0, "reward": 7.90295639038086, "reward_std": 0.3305318631231785, "rewards/combined_reward/mean": 7.902956199645996, "rewards/combined_reward/std": 2.215940606594086, "step": 49050 }, { "clip_ratio/high_max": 0.002650785709071594, "clip_ratio/high_mean": 0.0014853802326494285, "clip_ratio/low_mean": 0.0009712393534452228, "clip_ratio/low_min": 0.00028190591765451243, "clip_ratio/region_mean": 0.0024566195977968166, "completions/clipped_ratio": 0.0, "completions/max_length": 534.1, "completions/max_terminated_length": 534.1, "completions/mean_length": 210.8119857788086, "completions/mean_terminated_length": 210.8119857788086, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.031565971206873655, "epoch": 30.037343128252218, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5023319649.0, "reward": 8.121076679229736, "reward_std": 0.335410450398922, "rewards/combined_reward/mean": 8.121076202392578, "rewards/combined_reward/std": 2.368129003047943, "step": 49080 }, { "clip_ratio/high_max": 0.002809594735784534, "clip_ratio/high_mean": 0.0015829181282848972, "clip_ratio/low_mean": 0.0009456753524621793, "clip_ratio/low_min": 0.0002608813367260154, "clip_ratio/region_mean": 0.0025285934748050447, "completions/clipped_ratio": 0.0, "completions/max_length": 573.7, "completions/max_terminated_length": 573.7, "completions/mean_length": 204.9830780029297, "completions/mean_terminated_length": 204.9830780029297, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027759840494642654, "epoch": 30.055708601163147, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5026406944.0, "reward": 7.938922977447509, "reward_std": 0.31964071393013, "rewards/combined_reward/mean": 7.938922595977783, "rewards/combined_reward/std": 2.4540342807769777, "step": 49110 }, { "clip_ratio/high_max": 0.002161672102132191, "clip_ratio/high_mean": 0.0012961061428844308, "clip_ratio/low_mean": 0.0010585384282118563, "clip_ratio/low_min": 0.00032395093100300677, "clip_ratio/region_mean": 0.0023546445649117232, "completions/clipped_ratio": 0.0, "completions/max_length": 581.4, "completions/max_terminated_length": 581.4, "completions/mean_length": 208.36119995117187, "completions/mean_terminated_length": 208.36119995117187, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.02892414846767982, "epoch": 30.074074074074073, "frac_reward_zero_std": 0.14583333693444728, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5029507211.0, "reward": 8.176601028442382, "reward_std": 0.26506445109844207, "rewards/combined_reward/mean": 8.176600742340089, "rewards/combined_reward/std": 2.129307472705841, "step": 49140 }, { "clip_ratio/high_max": 0.002303638952920058, "clip_ratio/high_mean": 0.0013378329678137865, "clip_ratio/low_mean": 0.0009186032321546615, "clip_ratio/low_min": 0.00029021615506887125, "clip_ratio/region_mean": 0.0022564361987557884, "completions/clipped_ratio": 0.0, "completions/max_length": 547.8, "completions/max_terminated_length": 547.8, "completions/mean_length": 201.19375305175782, "completions/mean_terminated_length": 201.19375305175782, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.02763178825067977, "epoch": 30.092439546985002, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5032579955.0, "reward": 7.901103878021241, "reward_std": 0.31097846627235415, "rewards/combined_reward/mean": 7.901103782653808, "rewards/combined_reward/std": 2.491358518600464, "step": 49170 }, { "clip_ratio/high_max": 0.002269541236455552, "clip_ratio/high_mean": 0.0013093561465211678, "clip_ratio/low_mean": 0.0008912520343680323, "clip_ratio/low_min": 0.00016540944028141288, "clip_ratio/region_mean": 0.002200608163548168, "completions/clipped_ratio": 0.0, "completions/max_length": 551.9, "completions/max_terminated_length": 551.9, "completions/mean_length": 196.27318115234374, "completions/mean_terminated_length": 196.27318115234374, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.027055844695617757, "epoch": 30.110805019895928, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.007720947265625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5035633804.0, "reward": 8.26748080253601, "reward_std": 0.3553659781813622, "rewards/combined_reward/mean": 8.267480659484864, "rewards/combined_reward/std": 2.018851673603058, "step": 49200 }, { "clip_ratio/high_max": 0.002068494284079255, "clip_ratio/high_mean": 0.001212121988889218, "clip_ratio/low_mean": 0.0008291266989544965, "clip_ratio/low_min": 0.0002478680929925758, "clip_ratio/region_mean": 0.002041248685660927, "completions/clipped_ratio": 0.0, "completions/max_length": 601.7, "completions/max_terminated_length": 601.7, "completions/mean_length": 215.64792175292968, "completions/mean_terminated_length": 215.64792175292968, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.025377555439869564, "epoch": 30.129170492806857, "frac_reward_zero_std": 0.15833333730697632, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5038762052.0, "reward": 8.337560987472534, "reward_std": 0.3145431652665138, "rewards/combined_reward/mean": 8.337560892105103, "rewards/combined_reward/std": 2.004012906551361, "step": 49230 }, { "clip_ratio/high_max": 0.002558881374231229, "clip_ratio/high_mean": 0.0013855122831349339, "clip_ratio/low_mean": 0.0010276427399124562, "clip_ratio/low_min": 0.00029182632351876236, "clip_ratio/region_mean": 0.002413155027412965, "completions/clipped_ratio": 0.0, "completions/max_length": 683.5, "completions/max_terminated_length": 683.5, "completions/mean_length": 197.75781860351563, "completions/mean_terminated_length": 197.75781860351563, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.03135054669498156, "epoch": 30.147535965717783, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5041821602.0, "reward": 8.079259347915649, "reward_std": 0.3433577597141266, "rewards/combined_reward/mean": 8.079259204864503, "rewards/combined_reward/std": 2.165983247756958, "step": 49260 }, { "clip_ratio/high_max": 0.00267669243257842, "clip_ratio/high_mean": 0.0015927320785219005, "clip_ratio/low_mean": 0.0011275444864926006, "clip_ratio/low_min": 0.00035325609424035064, "clip_ratio/region_mean": 0.002720276562225384, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 199.66224517822266, "completions/mean_terminated_length": 199.66224517822266, "completions/min_length": 47.1, "completions/min_terminated_length": 47.1, "entropy": 0.03575247049642106, "epoch": 30.165901438628712, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.026611328125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 5044888465.0, "reward": 7.9243433475494385, "reward_std": 0.358055853843689, "rewards/combined_reward/mean": 7.924343013763428, "rewards/combined_reward/std": 2.4112816572189333, "step": 49290 }, { "clip_ratio/high_max": 0.0025776824496764068, "clip_ratio/high_mean": 0.0014832025837677066, "clip_ratio/low_mean": 0.0009549744019750506, "clip_ratio/low_min": 0.00028597948548849674, "clip_ratio/region_mean": 0.0024381769733736293, "completions/clipped_ratio": 0.0, "completions/max_length": 547.5, "completions/max_terminated_length": 547.5, "completions/mean_length": 185.5940155029297, "completions/mean_terminated_length": 185.5940155029297, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02817634295206517, "epoch": 30.184266911539638, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5047901306.0, "reward": 7.791605472564697, "reward_std": 0.30238020569086077, "rewards/combined_reward/mean": 7.7916053295135494, "rewards/combined_reward/std": 2.6432527184486387, "step": 49320 }, { "clip_ratio/high_max": 0.002483495081833098, "clip_ratio/high_mean": 0.001302430917348829, "clip_ratio/low_mean": 0.0008740243919116135, "clip_ratio/low_min": 0.00011693782483537992, "clip_ratio/region_mean": 0.0021764553069563894, "completions/clipped_ratio": 0.0, "completions/max_length": 593.6, "completions/max_terminated_length": 593.6, "completions/mean_length": 203.1627639770508, "completions/mean_terminated_length": 203.1627639770508, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.02675569918161879, "epoch": 30.202632384450567, "frac_reward_zero_std": 0.18750000484287738, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5050981611.0, "reward": 8.045748567581176, "reward_std": 0.37977737188339233, "rewards/combined_reward/mean": 8.045748376846314, "rewards/combined_reward/std": 2.398674488067627, "step": 49350 }, { "clip_ratio/high_max": 0.002278808406360137, "clip_ratio/high_mean": 0.0013318058833344064, "clip_ratio/low_mean": 0.0009899188259926935, "clip_ratio/low_min": 0.0003125613817246631, "clip_ratio/region_mean": 0.0023217247129650787, "completions/clipped_ratio": 0.0, "completions/max_length": 550.3, "completions/max_terminated_length": 550.3, "completions/mean_length": 203.84115142822264, "completions/mean_terminated_length": 203.84115142822264, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.026442512543871997, "epoch": 30.220997857361493, "frac_reward_zero_std": 0.18750000782310963, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5054064521.0, "reward": 8.04538335800171, "reward_std": 0.24030884206295014, "rewards/combined_reward/mean": 8.04538345336914, "rewards/combined_reward/std": 2.2990617513656617, "step": 49380 }, { "clip_ratio/high_max": 0.0021515603007477087, "clip_ratio/high_mean": 0.0012540651045128471, "clip_ratio/low_mean": 0.0008267713004897814, "clip_ratio/low_min": 0.00020142151576389247, "clip_ratio/region_mean": 0.0020808363944524897, "completions/clipped_ratio": 0.0, "completions/max_length": 580.6, "completions/max_terminated_length": 580.6, "completions/mean_length": 201.97656707763673, "completions/mean_terminated_length": 201.97656707763673, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028103222201267878, "epoch": 30.239363330272422, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5057140271.0, "reward": 8.127483987808228, "reward_std": 0.27579435557126997, "rewards/combined_reward/mean": 8.127483749389649, "rewards/combined_reward/std": 2.122669792175293, "step": 49410 }, { "clip_ratio/high_max": 0.002301257415577614, "clip_ratio/high_mean": 0.0013916450727265329, "clip_ratio/low_mean": 0.001095204084655658, "clip_ratio/low_min": 0.0003812992186188543, "clip_ratio/region_mean": 0.0024868491610201696, "completions/clipped_ratio": 0.0, "completions/max_length": 613.1, "completions/max_terminated_length": 613.1, "completions/mean_length": 195.91849365234376, "completions/mean_terminated_length": 195.91849365234376, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.02895507016219199, "epoch": 30.257728803183348, "frac_reward_zero_std": 0.15416667237877846, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5060192758.0, "reward": 8.127609729766846, "reward_std": 0.366756409406662, "rewards/combined_reward/mean": 8.127609539031983, "rewards/combined_reward/std": 2.1581670880317687, "step": 49440 }, { "clip_ratio/high_max": 0.0021608350580208935, "clip_ratio/high_mean": 0.0012416135300251578, "clip_ratio/low_mean": 0.0008407682258014877, "clip_ratio/low_min": 0.0002562286557804327, "clip_ratio/region_mean": 0.002082381781656295, "completions/clipped_ratio": 0.0, "completions/max_length": 596.4, "completions/max_terminated_length": 596.4, "completions/mean_length": 204.21250762939454, "completions/mean_terminated_length": 204.21250762939454, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.027052104690422615, "epoch": 30.276094276094277, "frac_reward_zero_std": 0.14583333544433116, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5063277094.0, "reward": 8.292634630203247, "reward_std": 0.3321470618247986, "rewards/combined_reward/mean": 8.292634344100952, "rewards/combined_reward/std": 1.9956907987594605, "step": 49470 }, { "clip_ratio/high_max": 0.002862721762115446, "clip_ratio/high_mean": 0.0015544105420606986, "clip_ratio/low_mean": 0.0010446580034719469, "clip_ratio/low_min": 0.000345929338800488, "clip_ratio/region_mean": 0.0025990685462602414, "completions/clipped_ratio": 0.0, "completions/max_length": 579.1, "completions/max_terminated_length": 579.1, "completions/mean_length": 197.16797332763673, "completions/mean_terminated_length": 197.16797332763673, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.029976111402114234, "epoch": 30.294459749005203, "frac_reward_zero_std": 0.1833333384245634, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5066334379.0, "reward": 8.305308532714843, "reward_std": 0.3572581112384796, "rewards/combined_reward/mean": 8.305308198928833, "rewards/combined_reward/std": 1.95100394487381, "step": 49500 }, { "clip_ratio/high_max": 0.0021992377519685154, "clip_ratio/high_mean": 0.0012625540416289975, "clip_ratio/low_mean": 0.0009181001971834727, "clip_ratio/low_min": 0.00025847690000470417, "clip_ratio/region_mean": 0.002180654225230683, "completions/clipped_ratio": 0.0, "completions/max_length": 608.7, "completions/max_terminated_length": 608.7, "completions/mean_length": 199.89844207763673, "completions/mean_terminated_length": 199.89844207763673, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02664963562662403, "epoch": 30.312825221916132, "frac_reward_zero_std": 0.1791666705161333, "grad_norm": 0.0242919921875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5069402149.0, "reward": 8.085855531692506, "reward_std": 0.338210241496563, "rewards/combined_reward/mean": 8.085855150222779, "rewards/combined_reward/std": 2.1163590788841247, "step": 49530 }, { "clip_ratio/high_max": 0.003005934669151126, "clip_ratio/high_mean": 0.0015528630957608888, "clip_ratio/low_mean": 0.0008882218064779105, "clip_ratio/low_min": 0.00025907179612355926, "clip_ratio/region_mean": 0.0024410848947203096, "completions/clipped_ratio": 0.0, "completions/max_length": 558.9, "completions/max_terminated_length": 558.9, "completions/mean_length": 203.46875457763673, "completions/mean_terminated_length": 203.46875457763673, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.03184774979017675, "epoch": 30.331190694827058, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 5072483629.0, "reward": 8.004020404815673, "reward_std": 0.2956389933824539, "rewards/combined_reward/mean": 8.004020071029663, "rewards/combined_reward/std": 2.471505641937256, "step": 49560 }, { "clip_ratio/high_max": 0.0022755569841441075, "clip_ratio/high_mean": 0.0012916705070286601, "clip_ratio/low_mean": 0.0009003379801773311, "clip_ratio/low_min": 0.00023356620101064134, "clip_ratio/region_mean": 0.0021920084807788955, "completions/clipped_ratio": 0.0, "completions/max_length": 535.8, "completions/max_terminated_length": 535.8, "completions/mean_length": 199.09141235351564, "completions/mean_terminated_length": 199.09141235351564, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.026498886725554862, "epoch": 30.349556167737987, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5075548300.0, "reward": 7.962245941162109, "reward_std": 0.3493009328842163, "rewards/combined_reward/mean": 7.962246036529541, "rewards/combined_reward/std": 2.402602255344391, "step": 49590 }, { "clip_ratio/high_max": 0.0028075449755609345, "clip_ratio/high_mean": 0.0015305013218797588, "clip_ratio/low_mean": 0.000919274063926423, "clip_ratio/low_min": 0.00020313035129220224, "clip_ratio/region_mean": 0.002449775357187415, "completions/clipped_ratio": 0.0, "completions/max_length": 621.4, "completions/max_terminated_length": 621.4, "completions/mean_length": 204.3541717529297, "completions/mean_terminated_length": 204.3541717529297, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03082981025800109, "epoch": 30.367921640648913, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5078633180.0, "reward": 8.289289903640746, "reward_std": 0.3411175191402435, "rewards/combined_reward/mean": 8.289289665222167, "rewards/combined_reward/std": 2.01280882358551, "step": 49620 }, { "clip_ratio/high_max": 0.0020786520753366252, "clip_ratio/high_mean": 0.0012248095527562933, "clip_ratio/low_mean": 0.000954253842428443, "clip_ratio/low_min": 0.0002971289320460831, "clip_ratio/region_mean": 0.002179063418589067, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 203.36406555175782, "completions/mean_terminated_length": 203.36406555175782, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.026572652944984534, "epoch": 30.386287113559842, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5081714258.0, "reward": 8.167085933685303, "reward_std": 0.31388905495405195, "rewards/combined_reward/mean": 8.167085790634156, "rewards/combined_reward/std": 2.0640917897224424, "step": 49650 }, { "clip_ratio/high_max": 0.0025717190651145454, "clip_ratio/high_mean": 0.0015685731309834712, "clip_ratio/low_mean": 0.0010799671177665004, "clip_ratio/low_min": 0.00034819327726533326, "clip_ratio/region_mean": 0.0026485402360170458, "completions/clipped_ratio": 0.0, "completions/max_length": 709.4, "completions/max_terminated_length": 709.4, "completions/mean_length": 206.75573272705077, "completions/mean_terminated_length": 206.75573272705077, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.0301145743423452, "epoch": 30.404652586470768, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5084808360.0, "reward": 7.959927034378052, "reward_std": 0.3143355369567871, "rewards/combined_reward/mean": 7.959926795959473, "rewards/combined_reward/std": 2.3788756251335146, "step": 49680 }, { "clip_ratio/high_max": 0.0020634865349469086, "clip_ratio/high_mean": 0.0011637754950546272, "clip_ratio/low_mean": 0.0008908381091411381, "clip_ratio/low_min": 0.00022103129958850332, "clip_ratio/region_mean": 0.0020546135870972647, "completions/clipped_ratio": 0.0, "completions/max_length": 592.6, "completions/max_terminated_length": 592.6, "completions/mean_length": 203.06588897705078, "completions/mean_terminated_length": 203.06588897705078, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.027525988686829806, "epoch": 30.423018059381697, "frac_reward_zero_std": 0.18333333991467954, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5087888293.0, "reward": 7.933100700378418, "reward_std": 0.3513632670044899, "rewards/combined_reward/mean": 7.933100461959839, "rewards/combined_reward/std": 2.440086233615875, "step": 49710 }, { "clip_ratio/high_max": 0.0019995573122287167, "clip_ratio/high_mean": 0.0011204301654894758, "clip_ratio/low_mean": 0.0008951450871488002, "clip_ratio/low_min": 0.00023451932550718388, "clip_ratio/region_mean": 0.002015575243907127, "completions/clipped_ratio": 0.0, "completions/max_length": 510.5, "completions/max_terminated_length": 510.5, "completions/mean_length": 184.57162017822264, "completions/mean_terminated_length": 184.57162017822264, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.025878249644301833, "epoch": 30.441383532292623, "frac_reward_zero_std": 0.21666667535901069, "grad_norm": 0.006072998046875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5090897208.0, "reward": 8.311819076538086, "reward_std": 0.24333405047655104, "rewards/combined_reward/mean": 8.31181879043579, "rewards/combined_reward/std": 2.0762124061584473, "step": 49740 }, { "clip_ratio/high_max": 0.002327863725561959, "clip_ratio/high_mean": 0.0012685269650925572, "clip_ratio/low_mean": 0.0009169249218151284, "clip_ratio/low_min": 0.0002794702037742051, "clip_ratio/region_mean": 0.002185451902914792, "completions/clipped_ratio": 0.0, "completions/max_length": 653.4, "completions/max_terminated_length": 653.4, "completions/mean_length": 208.28411865234375, "completions/mean_terminated_length": 208.28411865234375, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 0.029556687797109286, "epoch": 30.45974900520355, "frac_reward_zero_std": 0.1791666727513075, "grad_norm": 0.02880859375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5093997179.0, "reward": 8.094330883026123, "reward_std": 0.34350637942552564, "rewards/combined_reward/mean": 8.09433069229126, "rewards/combined_reward/std": 2.3767758131027223, "step": 49770 }, { "clip_ratio/high_max": 0.002265883347475513, "clip_ratio/high_mean": 0.0013130157666940553, "clip_ratio/low_mean": 0.0008829821385006653, "clip_ratio/low_min": 0.00018067496033230174, "clip_ratio/region_mean": 0.0021959978907640713, "completions/clipped_ratio": 0.0, "completions/max_length": 537.5, "completions/max_terminated_length": 537.5, "completions/mean_length": 201.44583740234376, "completions/mean_terminated_length": 201.44583740234376, "completions/min_length": 38.7, "completions/min_terminated_length": 38.7, "entropy": 0.027096401434391737, "epoch": 30.478114478114477, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.00494384765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5097070891.0, "reward": 8.058625841140747, "reward_std": 0.3102219939231873, "rewards/combined_reward/mean": 8.058625888824462, "rewards/combined_reward/std": 2.1948498487472534, "step": 49800 }, { "clip_ratio/high_max": 0.0023019703847239727, "clip_ratio/high_mean": 0.0012524643095578843, "clip_ratio/low_mean": 0.0010331321665338086, "clip_ratio/low_min": 0.00030825372862940034, "clip_ratio/region_mean": 0.002285596472938778, "completions/clipped_ratio": 0.0, "completions/max_length": 514.7, "completions/max_terminated_length": 514.7, "completions/mean_length": 202.3588623046875, "completions/mean_terminated_length": 202.3588623046875, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.028528757874543467, "epoch": 30.496479951025407, "frac_reward_zero_std": 0.18750000409781933, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5100148109.0, "reward": 7.989667987823486, "reward_std": 0.3318254932761192, "rewards/combined_reward/mean": 7.989667654037476, "rewards/combined_reward/std": 2.406808352470398, "step": 49830 }, { "clip_ratio/high_max": 0.0020844943127788914, "clip_ratio/high_mean": 0.0012090243469477476, "clip_ratio/low_mean": 0.0009477921286209797, "clip_ratio/low_min": 0.00029215966036038785, "clip_ratio/region_mean": 0.0021568164879378553, "completions/clipped_ratio": 0.0, "completions/max_length": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 196.8416717529297, "completions/mean_terminated_length": 196.8416717529297, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.025230163320278127, "epoch": 30.514845423936332, "frac_reward_zero_std": 0.1958333358168602, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5103204141.0, "reward": 8.132582473754884, "reward_std": 0.2812469467520714, "rewards/combined_reward/mean": 8.132582092285157, "rewards/combined_reward/std": 2.1426696181297302, "step": 49860 }, { "clip_ratio/high_max": 0.0024215982935857027, "clip_ratio/high_mean": 0.0014951407157544358, "clip_ratio/low_mean": 0.0009243345998887283, "clip_ratio/low_min": 0.00034223291659145615, "clip_ratio/region_mean": 0.002419475309337334, "completions/clipped_ratio": 0.0, "completions/max_length": 611.1, "completions/max_terminated_length": 611.1, "completions/mean_length": 213.16380615234374, "completions/mean_terminated_length": 213.16380615234374, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.028503892477601767, "epoch": 30.53321089684726, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5106322850.0, "reward": 8.049749040603638, "reward_std": 0.34275558590888977, "rewards/combined_reward/mean": 8.049748849868774, "rewards/combined_reward/std": 2.2740328192710875, "step": 49890 }, { "clip_ratio/high_max": 0.0022312981343323673, "clip_ratio/high_mean": 0.0013188261115525772, "clip_ratio/low_mean": 0.0010795137859531678, "clip_ratio/low_min": 0.0002815158516265607, "clip_ratio/region_mean": 0.0023983399121789263, "completions/clipped_ratio": 0.0, "completions/max_length": 636.2, "completions/max_terminated_length": 636.2, "completions/mean_length": 196.6510452270508, "completions/mean_terminated_length": 196.6510452270508, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.029636695282533765, "epoch": 30.551576369758187, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.01116943359375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5109378150.0, "reward": 8.0461856842041, "reward_std": 0.2703807935118675, "rewards/combined_reward/mean": 8.04618535041809, "rewards/combined_reward/std": 2.4137314438819883, "step": 49920 }, { "clip_ratio/high_max": 0.0022815189106040632, "clip_ratio/high_mean": 0.001334604781489664, "clip_ratio/low_mean": 0.0010087133102691344, "clip_ratio/low_min": 0.00026782531907277495, "clip_ratio/region_mean": 0.0023433180797534686, "completions/clipped_ratio": 0.0, "completions/max_length": 520.2, "completions/max_terminated_length": 520.2, "completions/mean_length": 192.5455795288086, "completions/mean_terminated_length": 192.5455795288086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.0286212877680858, "epoch": 30.569941842669117, "frac_reward_zero_std": 0.19583334177732467, "grad_norm": 0.0205078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5112417685.0, "reward": 8.01493434906006, "reward_std": 0.34977330565452575, "rewards/combined_reward/mean": 8.014934015274047, "rewards/combined_reward/std": 2.3103035569190977, "step": 49950 }, { "clip_ratio/high_max": 0.0024877374477606886, "clip_ratio/high_mean": 0.0014604606449817462, "clip_ratio/low_mean": 0.0008763482521317201, "clip_ratio/low_min": 0.00014718799308563273, "clip_ratio/region_mean": 0.0023368088877759874, "completions/clipped_ratio": 0.0, "completions/max_length": 582.4, "completions/max_terminated_length": 582.4, "completions/mean_length": 192.91146240234374, "completions/mean_terminated_length": 192.91146240234374, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029970465879887342, "epoch": 30.588307315580042, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5115458625.0, "reward": 7.953249979019165, "reward_std": 0.31676246747374537, "rewards/combined_reward/mean": 7.953249740600586, "rewards/combined_reward/std": 2.536316180229187, "step": 49980 }, { "clip_ratio/high_max": 0.0023967928107595073, "clip_ratio/high_mean": 0.0013650367414811625, "clip_ratio/low_mean": 0.0006893993850098923, "clip_ratio/low_min": 0.00019119827193208038, "clip_ratio/region_mean": 0.002054436126491055, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 190.17187881469727, "completions/mean_terminated_length": 190.17187881469727, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03497541495598853, "epoch": 30.606060606060606, "frac_reward_zero_std": 0.2395833395421505, "grad_norm": 0.0184326171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5118516215.0, "reward": 7.887003064155579, "reward_std": 0.33566857501864433, "rewards/combined_reward/mean": 7.887002944946289, "rewards/combined_reward/std": 2.4582499861717224, "step": 50010 }, { "clip_ratio/high_max": 0.001765499813579178, "clip_ratio/high_mean": 0.0010030081147609357, "clip_ratio/low_mean": 0.0006684006126306486, "clip_ratio/low_min": 0.0001734510539487625, "clip_ratio/region_mean": 0.0016714087318784247, "completions/clipped_ratio": 0.0, "completions/max_length": 553.6, "completions/max_terminated_length": 553.6, "completions/mean_length": 196.3307357788086, "completions/mean_terminated_length": 196.3307357788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.028249916108325123, "epoch": 30.624426078971535, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5121570285.0, "reward": 7.823524236679077, "reward_std": 0.3203813508152962, "rewards/combined_reward/mean": 7.823523998260498, "rewards/combined_reward/std": 2.447378134727478, "step": 50040 }, { "clip_ratio/high_max": 0.0018384619698432894, "clip_ratio/high_mean": 0.00104879817966624, "clip_ratio/low_mean": 0.0008155522285960614, "clip_ratio/low_min": 0.0002274053879470254, "clip_ratio/region_mean": 0.0018643504063220462, "completions/clipped_ratio": 0.0, "completions/max_length": 489.7, "completions/max_terminated_length": 489.7, "completions/mean_length": 196.51510772705078, "completions/mean_terminated_length": 196.51510772705078, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.02818445066610972, "epoch": 30.64279155188246, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5124625063.0, "reward": 8.047341108322144, "reward_std": 0.39276883006095886, "rewards/combined_reward/mean": 8.047340822219848, "rewards/combined_reward/std": 2.226548993587494, "step": 50070 }, { "clip_ratio/high_max": 0.0018492242040035005, "clip_ratio/high_mean": 0.001131082845434624, "clip_ratio/low_mean": 0.0008218456442894724, "clip_ratio/low_min": 0.00025420252083373874, "clip_ratio/region_mean": 0.0019529284860861178, "completions/clipped_ratio": 0.0, "completions/max_length": 648.2, "completions/max_terminated_length": 648.2, "completions/mean_length": 201.02943420410156, "completions/mean_terminated_length": 201.02943420410156, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.028470942145213486, "epoch": 30.66115702479339, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5127697176.0, "reward": 8.192980575561524, "reward_std": 0.33998923301696776, "rewards/combined_reward/mean": 8.192980575561524, "rewards/combined_reward/std": 2.2391059160232545, "step": 50100 }, { "clip_ratio/high_max": 0.001970926998668195, "clip_ratio/high_mean": 0.001028472622783738, "clip_ratio/low_mean": 0.0006212799529748736, "clip_ratio/low_min": 0.00015514614521331774, "clip_ratio/region_mean": 0.0016497525941910377, "completions/clipped_ratio": 0.0, "completions/max_length": 566.2, "completions/max_terminated_length": 566.2, "completions/mean_length": 186.1705795288086, "completions/mean_terminated_length": 186.1705795288086, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.027710301048743228, "epoch": 30.679522497704315, "frac_reward_zero_std": 0.22500000298023223, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5130712231.0, "reward": 7.96215238571167, "reward_std": 0.3398347467184067, "rewards/combined_reward/mean": 7.962152290344238, "rewards/combined_reward/std": 2.283978450298309, "step": 50130 }, { "clip_ratio/high_max": 0.0016982464410830289, "clip_ratio/high_mean": 0.0009150576542500251, "clip_ratio/low_mean": 0.0006742348678623482, "clip_ratio/low_min": 0.0001411233812784, "clip_ratio/region_mean": 0.0015892925160490752, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 191.77552642822266, "completions/mean_terminated_length": 191.77552642822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.027690796771397193, "epoch": 30.697887970615245, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.02880859375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5133748809.0, "reward": 8.080248832702637, "reward_std": 0.39809405505657197, "rewards/combined_reward/mean": 8.080248498916626, "rewards/combined_reward/std": 2.406204569339752, "step": 50160 }, { "clip_ratio/high_max": 0.0020229236358621466, "clip_ratio/high_mean": 0.0011780865124213355, "clip_ratio/low_mean": 0.0007248784547603767, "clip_ratio/low_min": 0.00019159243674948812, "clip_ratio/region_mean": 0.0019029649576017013, "completions/clipped_ratio": 0.0, "completions/max_length": 672.5, "completions/max_terminated_length": 672.5, "completions/mean_length": 206.7359420776367, "completions/mean_terminated_length": 206.7359420776367, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.02980406815186143, "epoch": 30.71625344352617, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.02587890625, "learning_rate": 5e-06, "loss": 0.0026, "num_tokens": 5136842835.0, "reward": 7.97392053604126, "reward_std": 0.38605008125305174, "rewards/combined_reward/mean": 7.973920392990112, "rewards/combined_reward/std": 2.4181426048278807, "step": 50190 }, { "clip_ratio/high_max": 0.0017234573470583806, "clip_ratio/high_mean": 0.000967604879042483, "clip_ratio/low_mean": 0.000755337090716542, "clip_ratio/low_min": 0.00016358382587592737, "clip_ratio/region_mean": 0.001722941969637759, "completions/clipped_ratio": 0.0, "completions/max_length": 532.6, "completions/max_terminated_length": 532.6, "completions/mean_length": 196.2979248046875, "completions/mean_terminated_length": 196.2979248046875, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.0261993963814651, "epoch": 30.7346189164371, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.00628662109375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5139896779.0, "reward": 7.92075481414795, "reward_std": 0.28286843523383143, "rewards/combined_reward/mean": 7.920754480361938, "rewards/combined_reward/std": 2.3961437463760378, "step": 50220 }, { "clip_ratio/high_max": 0.0017902854897935564, "clip_ratio/high_mean": 0.0010058408358114927, "clip_ratio/low_mean": 0.0007522921024550063, "clip_ratio/low_min": 0.00022626381566321168, "clip_ratio/region_mean": 0.001758132932203201, "completions/clipped_ratio": 0.0, "completions/max_length": 650.4, "completions/max_terminated_length": 650.4, "completions/mean_length": 209.91901702880858, "completions/mean_terminated_length": 209.91901702880858, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02759076973112921, "epoch": 30.752984389348025, "frac_reward_zero_std": 0.20833333991467953, "grad_norm": 0.0238037109375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5143003028.0, "reward": 8.276178693771362, "reward_std": 0.32518469393253324, "rewards/combined_reward/mean": 8.2761785030365, "rewards/combined_reward/std": 2.0016691744327546, "step": 50250 }, { "clip_ratio/high_max": 0.0019545879472085896, "clip_ratio/high_mean": 0.0011007992973948906, "clip_ratio/low_mean": 0.0008178013245924376, "clip_ratio/low_min": 0.0002802142516884487, "clip_ratio/region_mean": 0.001918600619925807, "completions/clipped_ratio": 0.0, "completions/max_length": 498.2, "completions/max_terminated_length": 498.2, "completions/mean_length": 186.34818420410156, "completions/mean_terminated_length": 186.34818420410156, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.027212341704095405, "epoch": 30.771349862258955, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.01214599609375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5146018765.0, "reward": 8.272121667861938, "reward_std": 0.2795271694660187, "rewards/combined_reward/mean": 8.27212119102478, "rewards/combined_reward/std": 2.0145677804946898, "step": 50280 }, { "clip_ratio/high_max": 0.0019784161466911126, "clip_ratio/high_mean": 0.0010572729004100741, "clip_ratio/low_mean": 0.0007841657371803497, "clip_ratio/low_min": 0.00021521065670337218, "clip_ratio/region_mean": 0.0018414386213407852, "completions/clipped_ratio": 0.0, "completions/max_length": 583.5, "completions/max_terminated_length": 583.5, "completions/mean_length": 201.0346420288086, "completions/mean_terminated_length": 201.0346420288086, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.028653032441313068, "epoch": 30.78971533516988, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5149090898.0, "reward": 8.063774108886719, "reward_std": 0.3123735450208187, "rewards/combined_reward/mean": 8.063773918151856, "rewards/combined_reward/std": 2.292705249786377, "step": 50310 }, { "clip_ratio/high_max": 0.0017988819861784577, "clip_ratio/high_mean": 0.0009910055482275008, "clip_ratio/low_mean": 0.0006931386744933358, "clip_ratio/low_min": 0.00017628209340424898, "clip_ratio/region_mean": 0.0016841442266013473, "completions/clipped_ratio": 0.0, "completions/max_length": 609.3, "completions/max_terminated_length": 609.3, "completions/mean_length": 208.50990142822266, "completions/mean_terminated_length": 208.50990142822266, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027469831208388012, "epoch": 30.80808080808081, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5152191736.0, "reward": 8.332897281646728, "reward_std": 0.3462861359119415, "rewards/combined_reward/mean": 8.33289680480957, "rewards/combined_reward/std": 1.9744460821151733, "step": 50340 }, { "clip_ratio/high_max": 0.00210918183195948, "clip_ratio/high_mean": 0.0012097852348233573, "clip_ratio/low_mean": 0.0007725863846038313, "clip_ratio/low_min": 0.00018931653855058054, "clip_ratio/region_mean": 0.00198237163858721, "completions/clipped_ratio": 0.0, "completions/max_length": 509.1, "completions/max_terminated_length": 509.1, "completions/mean_length": 198.61849517822264, "completions/mean_terminated_length": 198.61849517822264, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.027377524801219504, "epoch": 30.826446280991735, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5155254591.0, "reward": 7.99631085395813, "reward_std": 0.3289306879043579, "rewards/combined_reward/mean": 7.996310615539551, "rewards/combined_reward/std": 2.2763059735298157, "step": 50370 }, { "clip_ratio/high_max": 0.0019344224934078132, "clip_ratio/high_mean": 0.0011687240997465171, "clip_ratio/low_mean": 0.0010485024859614592, "clip_ratio/low_min": 0.00046148033798090184, "clip_ratio/region_mean": 0.0022172265892246895, "completions/clipped_ratio": 0.0, "completions/max_length": 658.7, "completions/max_terminated_length": 658.7, "completions/mean_length": 191.9752639770508, "completions/mean_terminated_length": 191.9752639770508, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.02982016356351475, "epoch": 30.84481175390266, "frac_reward_zero_std": 0.22916667461395263, "grad_norm": 0.006317138671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5158291936.0, "reward": 8.118744468688964, "reward_std": 0.2649808704853058, "rewards/combined_reward/mean": 8.11874418258667, "rewards/combined_reward/std": 2.1199942946434023, "step": 50400 }, { "clip_ratio/high_max": 0.0019373791976249777, "clip_ratio/high_mean": 0.0011577043953972558, "clip_ratio/low_mean": 0.0007059835830053392, "clip_ratio/low_min": 0.00019242617296792255, "clip_ratio/region_mean": 0.0018636879598489032, "completions/clipped_ratio": 0.0, "completions/max_length": 635.3, "completions/max_terminated_length": 635.3, "completions/mean_length": 205.50052642822266, "completions/mean_terminated_length": 205.50052642822266, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.03003380714605252, "epoch": 30.86317722681359, "frac_reward_zero_std": 0.16250000670552253, "grad_norm": 0.0294189453125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5161381218.0, "reward": 8.266984510421754, "reward_std": 0.3312294319272041, "rewards/combined_reward/mean": 8.266984367370606, "rewards/combined_reward/std": 2.053052020072937, "step": 50430 }, { "clip_ratio/high_max": 0.0022132205849629827, "clip_ratio/high_mean": 0.0012908625707495958, "clip_ratio/low_mean": 0.0009639537825326745, "clip_ratio/low_min": 0.00021379155684068488, "clip_ratio/region_mean": 0.002254816348431632, "completions/clipped_ratio": 0.0, "completions/max_length": 593.5, "completions/max_terminated_length": 593.5, "completions/mean_length": 193.15781860351564, "completions/mean_terminated_length": 193.15781860351564, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03124463784818848, "epoch": 30.88154269972452, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5164423104.0, "reward": 8.178910732269287, "reward_std": 0.3347243212163448, "rewards/combined_reward/mean": 8.178910160064698, "rewards/combined_reward/std": 2.156692457199097, "step": 50460 }, { "clip_ratio/high_max": 0.002130151646269951, "clip_ratio/high_mean": 0.001234873531939229, "clip_ratio/low_mean": 0.0007470041390964373, "clip_ratio/low_min": 0.00014653939651907422, "clip_ratio/region_mean": 0.0019818776951675923, "completions/clipped_ratio": 0.0, "completions/max_length": 559.2, "completions/max_terminated_length": 559.2, "completions/mean_length": 205.11016235351562, "completions/mean_terminated_length": 205.11016235351562, "completions/min_length": 45.1, "completions/min_terminated_length": 45.1, "entropy": 0.030816413182765245, "epoch": 30.899908172635445, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 5167510887.0, "reward": 8.064215707778931, "reward_std": 0.3839453250169754, "rewards/combined_reward/mean": 8.064215612411498, "rewards/combined_reward/std": 2.2279764890670775, "step": 50490 }, { "clip_ratio/high_max": 0.0019837357021363763, "clip_ratio/high_mean": 0.0012902724486290632, "clip_ratio/low_mean": 0.0006625766677946861, "clip_ratio/low_min": 0.00016824093933488863, "clip_ratio/region_mean": 0.001952849102478164, "completions/clipped_ratio": 0.0, "completions/max_length": 546.6, "completions/max_terminated_length": 546.6, "completions/mean_length": 198.20365142822266, "completions/mean_terminated_length": 198.20365142822266, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.030671873471389215, "epoch": 30.91827364554637, "frac_reward_zero_std": 0.23333334177732468, "grad_norm": 0.00653076171875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5170572149.0, "reward": 8.068662309646607, "reward_std": 0.27431790977716447, "rewards/combined_reward/mean": 8.06866192817688, "rewards/combined_reward/std": 2.39390230178833, "step": 50520 }, { "clip_ratio/high_max": 0.0017439026468006583, "clip_ratio/high_mean": 0.001031572236630988, "clip_ratio/low_mean": 0.0007617332502074229, "clip_ratio/low_min": 0.00025304915325250474, "clip_ratio/region_mean": 0.001793305495812092, "completions/clipped_ratio": 0.0, "completions/max_length": 539.8, "completions/max_terminated_length": 539.8, "completions/mean_length": 183.33021087646483, "completions/mean_terminated_length": 183.33021087646483, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.026964453095570207, "epoch": 30.9366391184573, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.0225830078125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5173576297.0, "reward": 7.772503423690796, "reward_std": 0.31927889585494995, "rewards/combined_reward/mean": 7.772503423690796, "rewards/combined_reward/std": 2.5467823147773743, "step": 50550 }, { "clip_ratio/high_max": 0.0020790488003209853, "clip_ratio/high_mean": 0.00129943901902152, "clip_ratio/low_mean": 0.0007501340189871068, "clip_ratio/low_min": 0.00024729119177209214, "clip_ratio/region_mean": 0.002049573047164207, "completions/clipped_ratio": 0.0, "completions/max_length": 630.4, "completions/max_terminated_length": 630.4, "completions/mean_length": 207.70781860351562, "completions/mean_terminated_length": 207.70781860351562, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.03057291715716322, "epoch": 30.955004591368226, "frac_reward_zero_std": 0.12500000298023223, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5176674055.0, "reward": 8.121082639694214, "reward_std": 0.3706978276371956, "rewards/combined_reward/mean": 8.121082496643066, "rewards/combined_reward/std": 2.042531394958496, "step": 50580 }, { "clip_ratio/high_max": 0.0020623749209335075, "clip_ratio/high_mean": 0.001110425896695233, "clip_ratio/low_mean": 0.0008115459579130402, "clip_ratio/low_min": 0.00019501788750252066, "clip_ratio/region_mean": 0.0019219718645520818, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 198.5093795776367, "completions/mean_terminated_length": 198.5093795776367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02853932452077667, "epoch": 30.973370064279155, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5179736491.0, "reward": 7.819536542892456, "reward_std": 0.3250934049487114, "rewards/combined_reward/mean": 7.819536352157593, "rewards/combined_reward/std": 2.554886794090271, "step": 50610 }, { "clip_ratio/high_max": 0.0016955786287629357, "clip_ratio/high_mean": 0.000983285645027839, "clip_ratio/low_mean": 0.000854139279666318, "clip_ratio/low_min": 0.00025125985800211005, "clip_ratio/region_mean": 0.001837424899955901, "completions/clipped_ratio": 0.0, "completions/max_length": 586.1, "completions/max_terminated_length": 586.1, "completions/mean_length": 203.4987030029297, "completions/mean_terminated_length": 203.4987030029297, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029447868978604675, "epoch": 30.99173553719008, "frac_reward_zero_std": 0.19166667014360428, "grad_norm": 0.01409912109375, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 5182818086.0, "reward": 7.975831699371338, "reward_std": 0.2867161944508553, "rewards/combined_reward/mean": 7.97583155632019, "rewards/combined_reward/std": 2.375467073917389, "step": 50640 }, { "clip_ratio/high_max": 0.0024415117179875673, "clip_ratio/high_mean": 0.0012024969424810031, "clip_ratio/low_mean": 0.0009493655967669272, "clip_ratio/low_min": 0.0003022068801820736, "clip_ratio/region_mean": 0.0021518625521299538, "completions/clipped_ratio": 0.0, "completions/max_length": 607.8, "completions/max_terminated_length": 607.8, "completions/mean_length": 206.91510772705078, "completions/mean_terminated_length": 206.91510772705078, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028409636915340774, "epoch": 31.010407101316193, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.01007080078125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5185912800.0, "reward": 8.094008445739746, "reward_std": 0.2679461777210236, "rewards/combined_reward/mean": 8.09400815963745, "rewards/combined_reward/std": 2.077799141407013, "step": 50670 }, { "clip_ratio/high_max": 0.0024063102711807003, "clip_ratio/high_mean": 0.0013847743341708944, "clip_ratio/low_mean": 0.0009585341248263527, "clip_ratio/low_min": 0.0002207871136003329, "clip_ratio/region_mean": 0.0023433084667582684, "completions/clipped_ratio": 0.0, "completions/max_length": 506.9, "completions/max_terminated_length": 506.9, "completions/mean_length": 202.59896392822264, "completions/mean_terminated_length": 202.59896392822264, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.029249098105356097, "epoch": 31.02877257422712, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5188990940.0, "reward": 8.079036664962768, "reward_std": 0.30800382792949677, "rewards/combined_reward/mean": 8.079036235809326, "rewards/combined_reward/std": 2.2495280146598815, "step": 50700 }, { "clip_ratio/high_max": 0.0028637835353341264, "clip_ratio/high_mean": 0.001586647786765146, "clip_ratio/low_mean": 0.0010197950616808763, "clip_ratio/low_min": 0.0003418221490089005, "clip_ratio/region_mean": 0.002606442831893219, "completions/clipped_ratio": 0.0, "completions/max_length": 610.8, "completions/max_terminated_length": 610.8, "completions/mean_length": 208.11823425292968, "completions/mean_terminated_length": 208.11823425292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.031200024446782968, "epoch": 31.04713804713805, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.029296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5192090274.0, "reward": 7.965782976150512, "reward_std": 0.3372292786836624, "rewards/combined_reward/mean": 7.9657830715179445, "rewards/combined_reward/std": 2.434370732307434, "step": 50730 }, { "clip_ratio/high_max": 0.002542599786829669, "clip_ratio/high_mean": 0.001508828791459867, "clip_ratio/low_mean": 0.001135143657666049, "clip_ratio/low_min": 0.00038316919138499844, "clip_ratio/region_mean": 0.002643972470347459, "completions/clipped_ratio": 0.0, "completions/max_length": 729.1, "completions/max_terminated_length": 729.1, "completions/mean_length": 206.56484832763672, "completions/mean_terminated_length": 206.56484832763672, "completions/min_length": 41.2, "completions/min_terminated_length": 41.2, "entropy": 0.03242469527758658, "epoch": 31.065503520048974, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0238037109375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5195183643.0, "reward": 8.085301876068115, "reward_std": 0.32649468779563906, "rewards/combined_reward/mean": 8.085301637649536, "rewards/combined_reward/std": 2.2464937806129455, "step": 50760 }, { "clip_ratio/high_max": 0.0026169307088518205, "clip_ratio/high_mean": 0.0015277212347427848, "clip_ratio/low_mean": 0.0010249487684632185, "clip_ratio/low_min": 0.0002895991732657421, "clip_ratio/region_mean": 0.002552669989624216, "completions/clipped_ratio": 0.0, "completions/max_length": 568.7, "completions/max_terminated_length": 568.7, "completions/mean_length": 207.5682373046875, "completions/mean_terminated_length": 207.5682373046875, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.029993926749254267, "epoch": 31.083868992959903, "frac_reward_zero_std": 0.20000000596046447, "grad_norm": 0.00469970703125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5198280865.0, "reward": 8.002644348144532, "reward_std": 0.27474926188588145, "rewards/combined_reward/mean": 8.002644109725953, "rewards/combined_reward/std": 2.3348461151123048, "step": 50790 }, { "clip_ratio/high_max": 0.0020122392428068756, "clip_ratio/high_mean": 0.0011696140395846062, "clip_ratio/low_mean": 0.0008788587671006099, "clip_ratio/low_min": 0.00022033717856781247, "clip_ratio/region_mean": 0.002048472820024472, "completions/clipped_ratio": 0.0, "completions/max_length": 564.6, "completions/max_terminated_length": 564.6, "completions/mean_length": 197.02292022705078, "completions/mean_terminated_length": 197.02292022705078, "completions/min_length": 41.3, "completions/min_terminated_length": 41.3, "entropy": 0.02515511658663551, "epoch": 31.10223446587083, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5201337593.0, "reward": 8.161256313323975, "reward_std": 0.3526189737021923, "rewards/combined_reward/mean": 8.1612557888031, "rewards/combined_reward/std": 2.239597570896149, "step": 50820 }, { "clip_ratio/high_max": 0.0025267612227859596, "clip_ratio/high_mean": 0.0013980192066810559, "clip_ratio/low_mean": 0.0009139611383943702, "clip_ratio/low_min": 0.0002552089303208049, "clip_ratio/region_mean": 0.0023119803566563255, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 202.77187957763672, "completions/mean_terminated_length": 202.77187957763672, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 0.02834618012420833, "epoch": 31.12059993878176, "frac_reward_zero_std": 0.17500000782310962, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5204416397.0, "reward": 8.320333528518677, "reward_std": 0.28032771944999696, "rewards/combined_reward/mean": 8.320333337783813, "rewards/combined_reward/std": 2.06919070482254, "step": 50850 }, { "clip_ratio/high_max": 0.0021413540108672652, "clip_ratio/high_mean": 0.001306658821098002, "clip_ratio/low_mean": 0.001090175514643003, "clip_ratio/low_min": 0.00033134306092203283, "clip_ratio/region_mean": 0.002396834357447612, "completions/clipped_ratio": 0.0, "completions/max_length": 591.6, "completions/max_terminated_length": 591.6, "completions/mean_length": 209.11875610351564, "completions/mean_terminated_length": 209.11875610351564, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.029239785702278216, "epoch": 31.138965411692684, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.013916015625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5207519573.0, "reward": 8.163656234741211, "reward_std": 0.33727664351463316, "rewards/combined_reward/mean": 8.163656187057494, "rewards/combined_reward/std": 2.0884928822517397, "step": 50880 }, { "clip_ratio/high_max": 0.0034976304547550776, "clip_ratio/high_mean": 0.0018976544170679214, "clip_ratio/low_mean": 0.0010793458926855238, "clip_ratio/low_min": 0.0003022222243695675, "clip_ratio/region_mean": 0.002977000271008971, "completions/clipped_ratio": 0.0, "completions/max_length": 735.2, "completions/max_terminated_length": 735.2, "completions/mean_length": 197.6799530029297, "completions/mean_terminated_length": 197.6799530029297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.036219180142506954, "epoch": 31.157330884603613, "frac_reward_zero_std": 0.18333333916962147, "grad_norm": 0.053955078125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5210578824.0, "reward": 8.115336465835572, "reward_std": 0.34321079403162, "rewards/combined_reward/mean": 8.11533613204956, "rewards/combined_reward/std": 2.1447800397872925, "step": 50910 }, { "clip_ratio/high_max": 0.0029292868227154636, "clip_ratio/high_mean": 0.0016735974117182196, "clip_ratio/low_mean": 0.0010323186873089677, "clip_ratio/low_min": 0.00026129958278033885, "clip_ratio/region_mean": 0.0027059160730762716, "completions/clipped_ratio": 0.0, "completions/max_length": 590.6, "completions/max_terminated_length": 590.6, "completions/mean_length": 193.0796905517578, "completions/mean_terminated_length": 193.0796905517578, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.033101281818623346, "epoch": 31.17569635751454, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.0068359375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5213620410.0, "reward": 7.596145915985107, "reward_std": 0.3525001496076584, "rewards/combined_reward/mean": 7.596145725250244, "rewards/combined_reward/std": 2.7334131836891173, "step": 50940 }, { "clip_ratio/high_max": 0.0019426863936435742, "clip_ratio/high_mean": 0.0010565998611127725, "clip_ratio/low_mean": 0.0009964323117552945, "clip_ratio/low_min": 0.00029888846353666546, "clip_ratio/region_mean": 0.0020530321777187054, "completions/clipped_ratio": 0.0, "completions/max_length": 583.9, "completions/max_terminated_length": 583.9, "completions/mean_length": 201.11224517822265, "completions/mean_terminated_length": 201.11224517822265, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.025935955904424192, "epoch": 31.19406183042547, "frac_reward_zero_std": 0.20833334177732468, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5216692841.0, "reward": 7.976831340789795, "reward_std": 0.3523029759526253, "rewards/combined_reward/mean": 7.976830959320068, "rewards/combined_reward/std": 2.6147417306900023, "step": 50970 }, { "clip_ratio/high_max": 0.002353170314017916, "clip_ratio/high_mean": 0.0012769775080717712, "clip_ratio/low_mean": 0.000888403413288567, "clip_ratio/low_min": 0.00021791596664115785, "clip_ratio/region_mean": 0.002165380912386657, "completions/clipped_ratio": 0.0, "completions/max_length": 544.6, "completions/max_terminated_length": 544.6, "completions/mean_length": 199.43959045410156, "completions/mean_terminated_length": 199.43959045410156, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.029019720464323957, "epoch": 31.212427303336394, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.0040283203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5219758849.0, "reward": 8.293182611465454, "reward_std": 0.30132745057344434, "rewards/combined_reward/mean": 8.293182134628296, "rewards/combined_reward/std": 2.0013824582099913, "step": 51000 }, { "clip_ratio/high_max": 0.002504993670542414, "clip_ratio/high_mean": 0.001333757893856576, "clip_ratio/low_mean": 0.0009279119628748352, "clip_ratio/low_min": 0.00023243624576328633, "clip_ratio/region_mean": 0.0022616698387234162, "completions/clipped_ratio": 0.0, "completions/max_length": 558.9, "completions/max_terminated_length": 558.9, "completions/mean_length": 200.6213607788086, "completions/mean_terminated_length": 200.6213607788086, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.027621515995512404, "epoch": 31.230792776247323, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.00714111328125, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 5222829395.0, "reward": 7.916784572601318, "reward_std": 0.24649427235126495, "rewards/combined_reward/mean": 7.916784286499023, "rewards/combined_reward/std": 2.3888877749443056, "step": 51030 }, { "clip_ratio/high_max": 0.0024772231074166482, "clip_ratio/high_mean": 0.0014682950556258827, "clip_ratio/low_mean": 0.0009345916760745847, "clip_ratio/low_min": 0.0002899419635165638, "clip_ratio/region_mean": 0.0024028867063558817, "completions/clipped_ratio": 0.0, "completions/max_length": 603.5, "completions/max_terminated_length": 603.5, "completions/mean_length": 199.90078735351562, "completions/mean_terminated_length": 199.90078735351562, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.029307618209471305, "epoch": 31.24915824915825, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.017578125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5225897174.0, "reward": 8.156802940368653, "reward_std": 0.3083158850669861, "rewards/combined_reward/mean": 8.15680284500122, "rewards/combined_reward/std": 2.1890612721443174, "step": 51060 }, { "clip_ratio/high_max": 0.0024408153234010874, "clip_ratio/high_mean": 0.0014338294458866585, "clip_ratio/low_mean": 0.0008428628529751829, "clip_ratio/low_min": 0.00019288260300527328, "clip_ratio/region_mean": 0.002276692305652735, "completions/clipped_ratio": 0.0, "completions/max_length": 559.6, "completions/max_terminated_length": 559.6, "completions/mean_length": 203.12240142822264, "completions/mean_terminated_length": 203.12240142822264, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.027880124157915513, "epoch": 31.26752372206918, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.01055908203125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5228977324.0, "reward": 8.350228071212769, "reward_std": 0.3619993507862091, "rewards/combined_reward/mean": 8.350227785110473, "rewards/combined_reward/std": 1.8629270553588868, "step": 51090 }, { "clip_ratio/high_max": 0.0021524307167661997, "clip_ratio/high_mean": 0.001205724466732742, "clip_ratio/low_mean": 0.0009075789645673163, "clip_ratio/low_min": 0.00022873250927659682, "clip_ratio/region_mean": 0.002113303442699059, "completions/clipped_ratio": 0.0, "completions/max_length": 524.2, "completions/max_terminated_length": 524.2, "completions/mean_length": 200.27422332763672, "completions/mean_terminated_length": 200.27422332763672, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.02705140917096287, "epoch": 31.285889194980104, "frac_reward_zero_std": 0.15833333879709244, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5232046537.0, "reward": 8.171413707733155, "reward_std": 0.387431275844574, "rewards/combined_reward/mean": 8.171413326263428, "rewards/combined_reward/std": 2.1202943325042725, "step": 51120 }, { "clip_ratio/high_max": 0.0019307659502374008, "clip_ratio/high_mean": 0.001088531241111923, "clip_ratio/low_mean": 0.0009162077671741524, "clip_ratio/low_min": 0.00020580474811140447, "clip_ratio/region_mean": 0.0020047390171384905, "completions/clipped_ratio": 0.0, "completions/max_length": 620.4, "completions/max_terminated_length": 620.4, "completions/mean_length": 189.9127639770508, "completions/mean_terminated_length": 189.9127639770508, "completions/min_length": 43.9, "completions/min_terminated_length": 43.9, "entropy": 0.025490847788751124, "epoch": 31.30425466789103, "frac_reward_zero_std": 0.22083334252238274, "grad_norm": 0.0289306640625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5235075962.0, "reward": 8.232779407501221, "reward_std": 0.3096757814288139, "rewards/combined_reward/mean": 8.232779264450073, "rewards/combined_reward/std": 2.031292366981506, "step": 51150 }, { "clip_ratio/high_max": 0.002578267561815058, "clip_ratio/high_mean": 0.0014401175549816495, "clip_ratio/low_mean": 0.0008313952587438204, "clip_ratio/low_min": 0.00026343575615707474, "clip_ratio/region_mean": 0.0022715128216077574, "completions/clipped_ratio": 0.0, "completions/max_length": 697.7, "completions/max_terminated_length": 697.7, "completions/mean_length": 203.6158905029297, "completions/mean_terminated_length": 203.6158905029297, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.03144978638738394, "epoch": 31.32262014080196, "frac_reward_zero_std": 0.16250000298023223, "grad_norm": 0.007598876953125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5238158007.0, "reward": 7.947073459625244, "reward_std": 0.354627688229084, "rewards/combined_reward/mean": 7.947073125839234, "rewards/combined_reward/std": 2.422483134269714, "step": 51180 }, { "clip_ratio/high_max": 0.002075299440912204, "clip_ratio/high_mean": 0.0010958004254158975, "clip_ratio/low_mean": 0.0010525521498493617, "clip_ratio/low_min": 0.000344864942356556, "clip_ratio/region_mean": 0.0021483525900597065, "completions/clipped_ratio": 0.0, "completions/max_length": 613.3, "completions/max_terminated_length": 613.3, "completions/mean_length": 206.64193267822264, "completions/mean_terminated_length": 206.64193267822264, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02698438858302931, "epoch": 31.340985613712885, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5241251672.0, "reward": 8.005533933639526, "reward_std": 0.3356526032090187, "rewards/combined_reward/mean": 8.00553364753723, "rewards/combined_reward/std": 2.506549668312073, "step": 51210 }, { "clip_ratio/high_max": 0.0026650153692268455, "clip_ratio/high_mean": 0.0015848700277274475, "clip_ratio/low_mean": 0.0009590617901267251, "clip_ratio/low_min": 0.0002996911224424063, "clip_ratio/region_mean": 0.002543931797845289, "completions/clipped_ratio": 0.0, "completions/max_length": 671.7, "completions/max_terminated_length": 671.7, "completions/mean_length": 209.63568115234375, "completions/mean_terminated_length": 209.63568115234375, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.030868181477611265, "epoch": 31.359351086623814, "frac_reward_zero_std": 0.13750000521540642, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5244356833.0, "reward": 8.056804180145264, "reward_std": 0.3854732930660248, "rewards/combined_reward/mean": 8.056803846359253, "rewards/combined_reward/std": 2.2217564940452577, "step": 51240 }, { "clip_ratio/high_max": 0.0025823900405763803, "clip_ratio/high_mean": 0.0014846218042900243, "clip_ratio/low_mean": 0.0009218360045148681, "clip_ratio/low_min": 0.00024706313561182467, "clip_ratio/region_mean": 0.00240645782808618, "completions/clipped_ratio": 0.0, "completions/max_length": 572.2, "completions/max_terminated_length": 572.2, "completions/mean_length": 201.45755767822266, "completions/mean_terminated_length": 201.45755767822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02882600963736574, "epoch": 31.37771655953474, "frac_reward_zero_std": 0.2000000074505806, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5247430590.0, "reward": 8.334228610992431, "reward_std": 0.28181798607110975, "rewards/combined_reward/mean": 8.334228134155273, "rewards/combined_reward/std": 1.8706765532493592, "step": 51270 }, { "clip_ratio/high_max": 0.0024840631716263792, "clip_ratio/high_mean": 0.001353173944274507, "clip_ratio/low_mean": 0.0008992700658078926, "clip_ratio/low_min": 0.0001629336184123531, "clip_ratio/region_mean": 0.0022524440011087183, "completions/clipped_ratio": 0.0, "completions/max_length": 627.1, "completions/max_terminated_length": 627.1, "completions/mean_length": 196.21875610351563, "completions/mean_terminated_length": 196.21875610351563, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.027525225390369692, "epoch": 31.39608203244567, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.03369140625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5250484230.0, "reward": 7.9729650020599365, "reward_std": 0.34743292182683944, "rewards/combined_reward/mean": 7.972964525222778, "rewards/combined_reward/std": 2.334001564979553, "step": 51300 }, { "clip_ratio/high_max": 0.0022716662346889886, "clip_ratio/high_mean": 0.0013051856984626889, "clip_ratio/low_mean": 0.0009040872972642925, "clip_ratio/low_min": 0.00022096792308730074, "clip_ratio/region_mean": 0.002209272954496555, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 210.62969207763672, "completions/mean_terminated_length": 210.62969207763672, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.028301118360832333, "epoch": 31.414447505356595, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.040771484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5253593208.0, "reward": 8.067792463302613, "reward_std": 0.29769744873046877, "rewards/combined_reward/mean": 8.067792272567749, "rewards/combined_reward/std": 2.224889028072357, "step": 51330 }, { "clip_ratio/high_max": 0.0022674020607761727, "clip_ratio/high_mean": 0.0012913905789901036, "clip_ratio/low_mean": 0.00091799422322462, "clip_ratio/low_min": 0.00036255874268438976, "clip_ratio/region_mean": 0.002209384798576745, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 191.6588592529297, "completions/mean_terminated_length": 191.6588592529297, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02785262488760054, "epoch": 31.432812978267524, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.01361083984375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5256629338.0, "reward": 8.093691444396972, "reward_std": 0.3062927782535553, "rewards/combined_reward/mean": 8.093691158294678, "rewards/combined_reward/std": 2.2928096771240236, "step": 51360 }, { "clip_ratio/high_max": 0.002080372524505947, "clip_ratio/high_mean": 0.0011925872085460773, "clip_ratio/low_mean": 0.0007403203732489297, "clip_ratio/low_min": 0.00015820528375722157, "clip_ratio/region_mean": 0.0019329075936790712, "completions/clipped_ratio": 0.0, "completions/max_length": 543.8, "completions/max_terminated_length": 543.8, "completions/mean_length": 202.21875457763673, "completions/mean_terminated_length": 202.21875457763673, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02557947099364052, "epoch": 31.45117845117845, "frac_reward_zero_std": 0.20000000558793546, "grad_norm": 0.01336669921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5259706018.0, "reward": 8.220023488998413, "reward_std": 0.31621444821357725, "rewards/combined_reward/mean": 8.220023202896119, "rewards/combined_reward/std": 2.1880836486816406, "step": 51390 }, { "clip_ratio/high_max": 0.002259703727516656, "clip_ratio/high_mean": 0.0012760935624100966, "clip_ratio/low_mean": 0.0008608253173103246, "clip_ratio/low_min": 0.00020144539206133535, "clip_ratio/region_mean": 0.0021369188631069846, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 200.46328582763672, "completions/mean_terminated_length": 200.46328582763672, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "entropy": 0.027827288877839845, "epoch": 31.46954392408938, "frac_reward_zero_std": 0.1875000074505806, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5262775957.0, "reward": 7.977412843704224, "reward_std": 0.352967357635498, "rewards/combined_reward/mean": 7.977412748336792, "rewards/combined_reward/std": 2.3814804911613465, "step": 51420 }, { "clip_ratio/high_max": 0.0028769485565135256, "clip_ratio/high_mean": 0.0015601377221173608, "clip_ratio/low_mean": 0.000846608063147869, "clip_ratio/low_min": 0.00022400360685423948, "clip_ratio/region_mean": 0.002406745762952293, "completions/clipped_ratio": 0.0, "completions/max_length": 625.9, "completions/max_terminated_length": 625.9, "completions/mean_length": 208.0901107788086, "completions/mean_terminated_length": 208.0901107788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03028249801136553, "epoch": 31.487909397000305, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.02734375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5265875183.0, "reward": 8.121842050552369, "reward_std": 0.34350003451108935, "rewards/combined_reward/mean": 8.121841859817504, "rewards/combined_reward/std": 2.1323785185813904, "step": 51450 }, { "clip_ratio/high_max": 0.002036132218684846, "clip_ratio/high_mean": 0.0011944731798090894, "clip_ratio/low_mean": 0.0007586359139774382, "clip_ratio/low_min": 0.00022018096642568709, "clip_ratio/region_mean": 0.0019531090872381659, "completions/clipped_ratio": 0.0, "completions/max_length": 518.3, "completions/max_terminated_length": 518.3, "completions/mean_length": 189.7033935546875, "completions/mean_terminated_length": 189.7033935546875, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.024346027011051773, "epoch": 31.506274869911234, "frac_reward_zero_std": 0.25000000968575475, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5268903804.0, "reward": 7.961567449569702, "reward_std": 0.3112759441137314, "rewards/combined_reward/mean": 7.961567163467407, "rewards/combined_reward/std": 2.561136078834534, "step": 51480 }, { "clip_ratio/high_max": 0.002214509009597047, "clip_ratio/high_mean": 0.0013699785803813333, "clip_ratio/low_mean": 0.0008896589915821096, "clip_ratio/low_min": 0.00023294612983590924, "clip_ratio/region_mean": 0.0022596375609282406, "completions/clipped_ratio": 0.0, "completions/max_length": 584.5, "completions/max_terminated_length": 584.5, "completions/mean_length": 207.94271392822264, "completions/mean_terminated_length": 207.94271392822264, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028191453870385887, "epoch": 31.52464034282216, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.006195068359375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5272002464.0, "reward": 8.195439338684082, "reward_std": 0.3089434042572975, "rewards/combined_reward/mean": 8.195439195632934, "rewards/combined_reward/std": 2.0322444438934326, "step": 51510 }, { "clip_ratio/high_max": 0.0023410803531684603, "clip_ratio/high_mean": 0.0013410330289237511, "clip_ratio/low_mean": 0.0011258081295333492, "clip_ratio/low_min": 0.0003483225770954353, "clip_ratio/region_mean": 0.0024668411429350575, "completions/clipped_ratio": 0.0, "completions/max_length": 642.9, "completions/max_terminated_length": 642.9, "completions/mean_length": 208.9635467529297, "completions/mean_terminated_length": 208.9635467529297, "completions/min_length": 46.7, "completions/min_terminated_length": 46.7, "entropy": 0.029745790207137666, "epoch": 31.54300581573309, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5275105044.0, "reward": 8.013617563247681, "reward_std": 0.273813484609127, "rewards/combined_reward/mean": 8.013617372512817, "rewards/combined_reward/std": 2.2816741824150086, "step": 51540 }, { "clip_ratio/high_max": 0.0027956593434889025, "clip_ratio/high_mean": 0.0015561516949674115, "clip_ratio/low_mean": 0.0010792165077873506, "clip_ratio/low_min": 0.0003491457335864349, "clip_ratio/region_mean": 0.002635368198874251, "completions/clipped_ratio": 0.0, "completions/max_length": 597.7, "completions/max_terminated_length": 597.7, "completions/mean_length": 195.95677642822267, "completions/mean_terminated_length": 195.95677642822267, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.0311898493518432, "epoch": 31.561371288644015, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.0098876953125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5278157678.0, "reward": 8.087547826766968, "reward_std": 0.34337639063596725, "rewards/combined_reward/mean": 8.087547397613525, "rewards/combined_reward/std": 2.3256460785865785, "step": 51570 }, { "clip_ratio/high_max": 0.0023596596457840253, "clip_ratio/high_mean": 0.0013912909975867175, "clip_ratio/low_mean": 0.0009154720528992281, "clip_ratio/low_min": 0.0002824431947374251, "clip_ratio/region_mean": 0.002306763049030754, "completions/clipped_ratio": 0.0, "completions/max_length": 520.7, "completions/max_terminated_length": 520.7, "completions/mean_length": 186.7697952270508, "completions/mean_terminated_length": 186.7697952270508, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.027763148319597045, "epoch": 31.579736761554944, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5281175034.0, "reward": 7.894017887115479, "reward_std": 0.2709657587110996, "rewards/combined_reward/mean": 7.894017601013184, "rewards/combined_reward/std": 2.5565593600273133, "step": 51600 }, { "clip_ratio/high_max": 0.002489340903897149, "clip_ratio/high_mean": 0.0014648778915822428, "clip_ratio/low_mean": 0.0010229044203394248, "clip_ratio/low_min": 0.00032499881732898456, "clip_ratio/region_mean": 0.0024877823210166148, "completions/clipped_ratio": 0.0, "completions/max_length": 619.6, "completions/max_terminated_length": 619.6, "completions/mean_length": 198.44349517822266, "completions/mean_terminated_length": 198.44349517822266, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.03072751169092953, "epoch": 31.59810223446587, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.0113525390625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5284237217.0, "reward": 8.07998447418213, "reward_std": 0.32236442863941195, "rewards/combined_reward/mean": 8.079984426498413, "rewards/combined_reward/std": 2.2839622616767885, "step": 51630 }, { "clip_ratio/high_max": 0.002408748987848715, "clip_ratio/high_mean": 0.0013096828030635759, "clip_ratio/low_mean": 0.0010250835124073394, "clip_ratio/low_min": 0.00028860363381681963, "clip_ratio/region_mean": 0.0023347663329332134, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 196.61068267822264, "completions/mean_terminated_length": 196.61068267822264, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.030243717258175214, "epoch": 31.6164677073768, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.021484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5287292362.0, "reward": 7.974103927612305, "reward_std": 0.3806535959243774, "rewards/combined_reward/mean": 7.97410364151001, "rewards/combined_reward/std": 2.1393246054649353, "step": 51660 }, { "clip_ratio/high_max": 0.002851612396868101, "clip_ratio/high_mean": 0.0015561056745355018, "clip_ratio/low_mean": 0.0010492018533113877, "clip_ratio/low_min": 0.0004052375734318048, "clip_ratio/region_mean": 0.0026053075217835914, "completions/clipped_ratio": 0.0, "completions/max_length": 583.6, "completions/max_terminated_length": 583.6, "completions/mean_length": 200.1424545288086, "completions/mean_terminated_length": 200.1424545288086, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.032311274739913645, "epoch": 31.634833180287725, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.01080322265625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5290361069.0, "reward": 8.097915315628052, "reward_std": 0.3576870605349541, "rewards/combined_reward/mean": 8.097915315628052, "rewards/combined_reward/std": 2.2489626526832582, "step": 51690 }, { "clip_ratio/high_max": 0.002634415484499186, "clip_ratio/high_mean": 0.001468605163730293, "clip_ratio/low_mean": 0.0010381870958857083, "clip_ratio/low_min": 0.00026507386307154473, "clip_ratio/region_mean": 0.002506792247489405, "completions/clipped_ratio": 0.0, "completions/max_length": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 210.46042022705078, "completions/mean_terminated_length": 210.46042022705078, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03183820004730175, "epoch": 31.653198653198654, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0223388671875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5293469397.0, "reward": 7.959287548065186, "reward_std": 0.3253690958023071, "rewards/combined_reward/mean": 7.959287166595459, "rewards/combined_reward/std": 2.3356019020080567, "step": 51720 }, { "clip_ratio/high_max": 0.0030531537927648363, "clip_ratio/high_mean": 0.0016214712217333727, "clip_ratio/low_mean": 0.0009874869470271126, "clip_ratio/low_min": 0.00032713521820066185, "clip_ratio/region_mean": 0.0026089581835549323, "completions/clipped_ratio": 0.0, "completions/max_length": 590.5, "completions/max_terminated_length": 590.5, "completions/mean_length": 199.71588897705078, "completions/mean_terminated_length": 199.71588897705078, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.033885263232514264, "epoch": 31.67156412610958, "frac_reward_zero_std": 0.16250000484287738, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0037, "num_tokens": 5296536466.0, "reward": 7.935325622558594, "reward_std": 0.3752021387219429, "rewards/combined_reward/mean": 7.93532543182373, "rewards/combined_reward/std": 2.3400726556777953, "step": 51750 }, { "clip_ratio/high_max": 0.002522902761605413, "clip_ratio/high_mean": 0.0013334499923075781, "clip_ratio/low_mean": 0.0009914796229471298, "clip_ratio/low_min": 0.00025248430344314933, "clip_ratio/region_mean": 0.0023249296097977395, "completions/clipped_ratio": 0.0, "completions/max_length": 517.1, "completions/max_terminated_length": 517.1, "completions/mean_length": 189.1658920288086, "completions/mean_terminated_length": 189.1658920288086, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.027766952733509244, "epoch": 31.68992959902051, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.03857421875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5299563023.0, "reward": 8.051232814788818, "reward_std": 0.28497018814086916, "rewards/combined_reward/mean": 8.05123233795166, "rewards/combined_reward/std": 2.182452750205994, "step": 51780 }, { "clip_ratio/high_max": 0.0024406100346823223, "clip_ratio/high_mean": 0.001279601440425419, "clip_ratio/low_mean": 0.001031050583878823, "clip_ratio/low_min": 0.000341230547928717, "clip_ratio/region_mean": 0.0023106520192717047, "completions/clipped_ratio": 0.0, "completions/max_length": 525.1, "completions/max_terminated_length": 525.1, "completions/mean_length": 197.83255767822266, "completions/mean_terminated_length": 197.83255767822266, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.029479487993133564, "epoch": 31.708295071931435, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.01806640625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5302622860.0, "reward": 8.059419107437133, "reward_std": 0.30763392299413683, "rewards/combined_reward/mean": 8.059419012069702, "rewards/combined_reward/std": 2.465474784374237, "step": 51810 }, { "clip_ratio/high_max": 0.002498346146118517, "clip_ratio/high_mean": 0.0014219254169195968, "clip_ratio/low_mean": 0.0008997006931167562, "clip_ratio/low_min": 0.00024979112907506836, "clip_ratio/region_mean": 0.0023216261232543427, "completions/clipped_ratio": 0.0, "completions/max_length": 595.4, "completions/max_terminated_length": 595.4, "completions/mean_length": 203.6463592529297, "completions/mean_terminated_length": 203.6463592529297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03187757823616266, "epoch": 31.726660544842364, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 5305705022.0, "reward": 7.921184968948364, "reward_std": 0.3878888189792633, "rewards/combined_reward/mean": 7.921184730529785, "rewards/combined_reward/std": 2.2424838304519654, "step": 51840 }, { "clip_ratio/high_max": 0.0028825590959362063, "clip_ratio/high_mean": 0.001610281935184806, "clip_ratio/low_mean": 0.0009695908919335731, "clip_ratio/low_min": 0.0002762857843966534, "clip_ratio/region_mean": 0.002579872830635092, "completions/clipped_ratio": 0.0, "completions/max_length": 664.7, "completions/max_terminated_length": 664.7, "completions/mean_length": 201.5862030029297, "completions/mean_terminated_length": 201.5862030029297, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03176914555951953, "epoch": 31.74502601775329, "frac_reward_zero_std": 0.17500000409781932, "grad_norm": 0.01495361328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5308779273.0, "reward": 7.763537359237671, "reward_std": 0.3207291439175606, "rewards/combined_reward/mean": 7.763536977767944, "rewards/combined_reward/std": 2.5863651990890504, "step": 51870 }, { "clip_ratio/high_max": 0.002551686686638277, "clip_ratio/high_mean": 0.0014085671178084643, "clip_ratio/low_mean": 0.0009991360890126088, "clip_ratio/low_min": 0.00023827365245476055, "clip_ratio/region_mean": 0.0024077032154309562, "completions/clipped_ratio": 0.0, "completions/max_length": 716.4, "completions/max_terminated_length": 716.4, "completions/mean_length": 201.2869842529297, "completions/mean_terminated_length": 201.2869842529297, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.030547129518042008, "epoch": 31.76339149066422, "frac_reward_zero_std": 0.21250000558793544, "grad_norm": 0.022216796875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5311852375.0, "reward": 8.103887224197388, "reward_std": 0.3185629591345787, "rewards/combined_reward/mean": 8.103886842727661, "rewards/combined_reward/std": 2.2610586285591125, "step": 51900 }, { "clip_ratio/high_max": 0.0024469334219854016, "clip_ratio/high_mean": 0.0013988873772177613, "clip_ratio/low_mean": 0.0009105137887672754, "clip_ratio/low_min": 0.0002768613708515962, "clip_ratio/region_mean": 0.0023094011475526107, "completions/clipped_ratio": 0.0, "completions/max_length": 615.8, "completions/max_terminated_length": 615.8, "completions/mean_length": 204.29922485351562, "completions/mean_terminated_length": 204.29922485351562, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.028652547563736638, "epoch": 31.781756963575145, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.0211181640625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5314937044.0, "reward": 8.065014553070068, "reward_std": 0.32940748929977415, "rewards/combined_reward/mean": 8.065014219284057, "rewards/combined_reward/std": 2.1767328679561615, "step": 51930 }, { "clip_ratio/high_max": 0.0023137409933648694, "clip_ratio/high_mean": 0.0013013996360920525, "clip_ratio/low_mean": 0.0010451341489291128, "clip_ratio/low_min": 0.00036301906717805345, "clip_ratio/region_mean": 0.0023465337998156127, "completions/clipped_ratio": 0.0, "completions/max_length": 570.6, "completions/max_terminated_length": 570.6, "completions/mean_length": 199.72266387939453, "completions/mean_terminated_length": 199.72266387939453, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.028151076023156443, "epoch": 31.800122436486074, "frac_reward_zero_std": 0.1583333358168602, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5318004139.0, "reward": 7.924777984619141, "reward_std": 0.366411592066288, "rewards/combined_reward/mean": 7.924777698516846, "rewards/combined_reward/std": 2.184461957216263, "step": 51960 }, { "clip_ratio/high_max": 0.0024923171586124226, "clip_ratio/high_mean": 0.0014215418426222943, "clip_ratio/low_mean": 0.0010132582003279822, "clip_ratio/low_min": 0.000358472569011307, "clip_ratio/region_mean": 0.0024348000219712655, "completions/clipped_ratio": 0.0, "completions/max_length": 549.1, "completions/max_terminated_length": 549.1, "completions/mean_length": 205.79531860351562, "completions/mean_terminated_length": 205.79531860351562, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.029297824886937936, "epoch": 31.818487909397, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.027587890625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5321094553.0, "reward": 8.064846515655518, "reward_std": 0.37850964069366455, "rewards/combined_reward/mean": 8.064846229553222, "rewards/combined_reward/std": 2.217173767089844, "step": 51990 }, { "clip_ratio/high_max": 0.0023543306349893102, "clip_ratio/high_mean": 0.0014004178359755316, "clip_ratio/low_mean": 0.0007936036814498948, "clip_ratio/low_min": 0.0002789978010696359, "clip_ratio/region_mean": 0.0021940215054200964, "completions/clipped_ratio": 0.0, "completions/max_length": 691.5714285714286, "completions/max_terminated_length": 691.5714285714286, "completions/mean_length": 195.63504900251115, "completions/mean_terminated_length": 195.63504900251115, "completions/min_length": 37.714285714285715, "completions/min_terminated_length": 37.714285714285715, "entropy": 0.031245385319925843, "epoch": 31.836241199877563, "frac_reward_zero_std": 0.20833333688122885, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5324171507.0, "reward": 7.942483084542411, "reward_std": 0.3419367415564401, "rewards/combined_reward/mean": 7.942482948303223, "rewards/combined_reward/std": 2.3056354352406094, "step": 52020 }, { "clip_ratio/high_max": 0.0017496044621414816, "clip_ratio/high_mean": 0.001087061113503296, "clip_ratio/low_mean": 0.0008223660358150179, "clip_ratio/low_min": 0.0003104349792314072, "clip_ratio/region_mean": 0.0019094271352514624, "completions/clipped_ratio": 0.0, "completions/max_length": 664.2, "completions/max_terminated_length": 664.2, "completions/mean_length": 200.35833892822265, "completions/mean_terminated_length": 200.35833892822265, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.02728079470495383, "epoch": 31.854606672788492, "frac_reward_zero_std": 0.20416667237877845, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5327241043.0, "reward": 8.240023517608643, "reward_std": 0.3286887094378471, "rewards/combined_reward/mean": 8.240022993087768, "rewards/combined_reward/std": 2.069802713394165, "step": 52050 }, { "clip_ratio/high_max": 0.0020716843345629362, "clip_ratio/high_mean": 0.0012531612007781708, "clip_ratio/low_mean": 0.0007912947387618866, "clip_ratio/low_min": 0.000210009917645948, "clip_ratio/region_mean": 0.0020444559331129616, "completions/clipped_ratio": 0.0, "completions/max_length": 623.4, "completions/max_terminated_length": 623.4, "completions/mean_length": 198.14635620117187, "completions/mean_terminated_length": 198.14635620117187, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03210041119406621, "epoch": 31.872972145699418, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.00836181640625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5330302085.0, "reward": 8.304969692230225, "reward_std": 0.3283235654234886, "rewards/combined_reward/mean": 8.304969692230225, "rewards/combined_reward/std": 1.9757665872573853, "step": 52080 }, { "clip_ratio/high_max": 0.0023516670024643343, "clip_ratio/high_mean": 0.0013014118155600348, "clip_ratio/low_mean": 0.0008861303442851446, "clip_ratio/low_min": 0.0002981821778424395, "clip_ratio/region_mean": 0.0021875421594207485, "completions/clipped_ratio": 0.0, "completions/max_length": 551.6, "completions/max_terminated_length": 551.6, "completions/mean_length": 195.58151397705078, "completions/mean_terminated_length": 195.58151397705078, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.030632948347677786, "epoch": 31.891337618610347, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.0084228515625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5333353278.0, "reward": 8.04578881263733, "reward_std": 0.3200660958886147, "rewards/combined_reward/mean": 8.045788764953613, "rewards/combined_reward/std": 2.269169783592224, "step": 52110 }, { "clip_ratio/high_max": 0.001983173711535831, "clip_ratio/high_mean": 0.0012086528816629045, "clip_ratio/low_mean": 0.0006501743669408218, "clip_ratio/low_min": 0.00018478812780813313, "clip_ratio/region_mean": 0.0018588272345368751, "completions/clipped_ratio": 0.0, "completions/max_length": 593.7, "completions/max_terminated_length": 593.7, "completions/mean_length": 205.00339050292968, "completions/mean_terminated_length": 205.00339050292968, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.031711028330028056, "epoch": 31.909703091521273, "frac_reward_zero_std": 0.19583333767950534, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5336440651.0, "reward": 8.177693223953247, "reward_std": 0.36001883149147035, "rewards/combined_reward/mean": 8.177692937850953, "rewards/combined_reward/std": 2.0722742915153503, "step": 52140 }, { "clip_ratio/high_max": 0.0020236131502315404, "clip_ratio/high_mean": 0.0012159189345159879, "clip_ratio/low_mean": 0.0007298031381651526, "clip_ratio/low_min": 0.00024145365653870007, "clip_ratio/region_mean": 0.0019457220691644276, "completions/clipped_ratio": 0.0, "completions/max_length": 568.1, "completions/max_terminated_length": 568.1, "completions/mean_length": 194.67682647705078, "completions/mean_terminated_length": 194.67682647705078, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.028990189839775363, "epoch": 31.928068564432202, "frac_reward_zero_std": 0.19583333879709244, "grad_norm": 0.037353515625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5339488370.0, "reward": 7.809457874298095, "reward_std": 0.37316847443580625, "rewards/combined_reward/mean": 7.809457731246948, "rewards/combined_reward/std": 2.6720035314559936, "step": 52170 }, { "clip_ratio/high_max": 0.0016268214167212136, "clip_ratio/high_mean": 0.0009784467426167491, "clip_ratio/low_mean": 0.0006973666687069151, "clip_ratio/low_min": 0.00017532422449827815, "clip_ratio/region_mean": 0.0016758134142340472, "completions/clipped_ratio": 0.0, "completions/max_length": 598.2, "completions/max_terminated_length": 598.2, "completions/mean_length": 191.0932357788086, "completions/mean_terminated_length": 191.0932357788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02784273320964227, "epoch": 31.946434037343128, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5342522328.0, "reward": 8.087463283538819, "reward_std": 0.28906250596046446, "rewards/combined_reward/mean": 8.087463092803954, "rewards/combined_reward/std": 2.1236982703208924, "step": 52200 }, { "clip_ratio/high_max": 0.0020248369915255656, "clip_ratio/high_mean": 0.0011382250573660712, "clip_ratio/low_mean": 0.0008620835797046311, "clip_ratio/low_min": 0.00025557956517635225, "clip_ratio/region_mean": 0.002000308623246383, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 209.26693115234374, "completions/mean_terminated_length": 209.26693115234374, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029252709657885134, "epoch": 31.964799510254057, "frac_reward_zero_std": 0.12916666977107524, "grad_norm": 0.00457763671875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5345626073.0, "reward": 7.855771827697754, "reward_std": 0.36372620463371275, "rewards/combined_reward/mean": 7.855771493911743, "rewards/combined_reward/std": 2.3729748487472535, "step": 52230 }, { "clip_ratio/high_max": 0.002119567827321589, "clip_ratio/high_mean": 0.0011385759292655469, "clip_ratio/low_mean": 0.0007776325277518481, "clip_ratio/low_min": 0.00022451866849830063, "clip_ratio/region_mean": 0.0019162084434356074, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 197.49219360351563, "completions/mean_terminated_length": 197.49219360351563, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.02987694162875414, "epoch": 31.983164983164983, "frac_reward_zero_std": 0.19166667498648166, "grad_norm": 0.015869140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5348684603.0, "reward": 8.017652654647828, "reward_std": 0.3025496557354927, "rewards/combined_reward/mean": 8.017652225494384, "rewards/combined_reward/std": 2.4249361991882323, "step": 52260 }, { "clip_ratio/high_max": 0.0019085495618217793, "clip_ratio/high_mean": 0.001081475094161149, "clip_ratio/low_mean": 0.0008062163280078866, "clip_ratio/low_min": 0.00014555403107746704, "clip_ratio/region_mean": 0.0018876914195449198, "completions/clipped_ratio": 0.0, "completions/max_length": 620.7, "completions/max_terminated_length": 620.7, "completions/mean_length": 203.97760925292968, "completions/mean_terminated_length": 203.97760925292968, "completions/min_length": 41.5, "completions/min_terminated_length": 41.5, "entropy": 0.027106446374329875, "epoch": 32.00183654729109, "frac_reward_zero_std": 0.12916666865348816, "grad_norm": 0.00848388671875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5351768037.0, "reward": 8.08692455291748, "reward_std": 0.3385659709572792, "rewards/combined_reward/mean": 8.086924505233764, "rewards/combined_reward/std": 2.2029404640197754, "step": 52290 }, { "clip_ratio/high_max": 0.0029337376768429142, "clip_ratio/high_mean": 0.001563225666419991, "clip_ratio/low_mean": 0.0008708216687712896, "clip_ratio/low_min": 0.00018922836364557345, "clip_ratio/region_mean": 0.0024340473256112697, "completions/clipped_ratio": 0.0, "completions/max_length": 616.5, "completions/max_terminated_length": 616.5, "completions/mean_length": 203.59219360351562, "completions/mean_terminated_length": 203.59219360351562, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.030299623621006806, "epoch": 32.02020202020202, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5354849991.0, "reward": 7.929837465286255, "reward_std": 0.27567662447690966, "rewards/combined_reward/mean": 7.929837226867676, "rewards/combined_reward/std": 2.1902472734451295, "step": 52320 }, { "clip_ratio/high_max": 0.0026586486662078337, "clip_ratio/high_mean": 0.0014206249054647439, "clip_ratio/low_mean": 0.0010765524029314596, "clip_ratio/low_min": 0.00037796297013604395, "clip_ratio/region_mean": 0.00249717731300431, "completions/clipped_ratio": 0.0, "completions/max_length": 547.7, "completions/max_terminated_length": 547.7, "completions/mean_length": 209.49115142822265, "completions/mean_terminated_length": 209.49115142822265, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.031379313316817085, "epoch": 32.03856749311295, "frac_reward_zero_std": 0.14583333805203438, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5357954597.0, "reward": 8.121289777755738, "reward_std": 0.3025962769985199, "rewards/combined_reward/mean": 8.121289587020874, "rewards/combined_reward/std": 2.3337199330329894, "step": 52350 }, { "clip_ratio/high_max": 0.0025333001406882736, "clip_ratio/high_mean": 0.0013756238615314942, "clip_ratio/low_mean": 0.0009421318097641536, "clip_ratio/low_min": 0.00027872461566100053, "clip_ratio/region_mean": 0.0023177556814819884, "completions/clipped_ratio": 0.0, "completions/max_length": 608.9, "completions/max_terminated_length": 608.9, "completions/mean_length": 205.0526092529297, "completions/mean_terminated_length": 205.0526092529297, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027326928762098152, "epoch": 32.05693296602387, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.0089111328125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5361042159.0, "reward": 7.915676641464233, "reward_std": 0.3603060394525528, "rewards/combined_reward/mean": 7.915676403045654, "rewards/combined_reward/std": 2.5026729702949524, "step": 52380 }, { "clip_ratio/high_max": 0.002251622829256424, "clip_ratio/high_mean": 0.0013453076559623394, "clip_ratio/low_mean": 0.0009445340781288299, "clip_ratio/low_min": 0.00028353661085323744, "clip_ratio/region_mean": 0.0022898417393056054, "completions/clipped_ratio": 0.0, "completions/max_length": 657.5, "completions/max_terminated_length": 657.5, "completions/mean_length": 207.31328582763672, "completions/mean_terminated_length": 207.31328582763672, "completions/min_length": 41.8, "completions/min_terminated_length": 41.8, "entropy": 0.03006405766742925, "epoch": 32.0752984389348, "frac_reward_zero_std": 0.1416666705161333, "grad_norm": 0.003265380859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5364138402.0, "reward": 8.188792371749878, "reward_std": 0.27933177873492243, "rewards/combined_reward/mean": 8.1887921333313, "rewards/combined_reward/std": 2.125291383266449, "step": 52410 }, { "clip_ratio/high_max": 0.002389416057364239, "clip_ratio/high_mean": 0.001318566526970244, "clip_ratio/low_mean": 0.0007932388781531093, "clip_ratio/low_min": 0.00023955206537114768, "clip_ratio/region_mean": 0.0021118054117929812, "completions/clipped_ratio": 0.0, "completions/max_length": 547.5, "completions/max_terminated_length": 547.5, "completions/mean_length": 200.75729522705078, "completions/mean_terminated_length": 200.75729522705078, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.027939929525988796, "epoch": 32.09366391184573, "frac_reward_zero_std": 0.21250000670552255, "grad_norm": 0.01025390625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5367209470.0, "reward": 7.905392742156982, "reward_std": 0.334275396168232, "rewards/combined_reward/mean": 7.905392694473266, "rewards/combined_reward/std": 2.496955990791321, "step": 52440 }, { "clip_ratio/high_max": 0.002280954030478218, "clip_ratio/high_mean": 0.0012408581607208666, "clip_ratio/low_mean": 0.0008969184704255895, "clip_ratio/low_min": 0.00025107956729092016, "clip_ratio/region_mean": 0.002137776637391653, "completions/clipped_ratio": 0.0, "completions/max_length": 568.8, "completions/max_terminated_length": 568.8, "completions/mean_length": 196.7296936035156, "completions/mean_terminated_length": 196.7296936035156, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.027011076593771575, "epoch": 32.11202938475666, "frac_reward_zero_std": 0.14583333879709243, "grad_norm": 0.0086669921875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5370265072.0, "reward": 8.279846286773681, "reward_std": 0.3571659713983536, "rewards/combined_reward/mean": 8.279846000671387, "rewards/combined_reward/std": 1.997243094444275, "step": 52470 }, { "clip_ratio/high_max": 0.0022330776089802386, "clip_ratio/high_mean": 0.0012952283621416427, "clip_ratio/low_mean": 0.0008027224379475228, "clip_ratio/low_min": 0.0002197480319106641, "clip_ratio/region_mean": 0.0020979507966937187, "completions/clipped_ratio": 0.0, "completions/max_length": 574.1, "completions/max_terminated_length": 574.1, "completions/mean_length": 214.7244842529297, "completions/mean_terminated_length": 214.7244842529297, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.02609666120260954, "epoch": 32.13039485766758, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.004058837890625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5373389774.0, "reward": 8.354581499099732, "reward_std": 0.301253741979599, "rewards/combined_reward/mean": 8.354581165313721, "rewards/combined_reward/std": 1.9739832758903504, "step": 52500 }, { "clip_ratio/high_max": 0.0026539842539932577, "clip_ratio/high_mean": 0.0015138678087775286, "clip_ratio/low_mean": 0.0010281925828773335, "clip_ratio/low_min": 0.00029159336457572257, "clip_ratio/region_mean": 0.0025420603958385374, "completions/clipped_ratio": 0.0, "completions/max_length": 649.4, "completions/max_terminated_length": 649.4, "completions/mean_length": 196.70287017822267, "completions/mean_terminated_length": 196.70287017822267, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03183686098394294, "epoch": 32.14876033057851, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5376445273.0, "reward": 8.082514238357543, "reward_std": 0.3109033301472664, "rewards/combined_reward/mean": 8.082513761520385, "rewards/combined_reward/std": 2.1543776035308837, "step": 52530 }, { "clip_ratio/high_max": 0.003301537070365157, "clip_ratio/high_mean": 0.0017927742364311902, "clip_ratio/low_mean": 0.0010652117636103262, "clip_ratio/low_min": 0.0002690169729855067, "clip_ratio/region_mean": 0.002857986019807868, "completions/clipped_ratio": 0.0, "completions/max_length": 779.2, "completions/max_terminated_length": 779.2, "completions/mean_length": 199.3820358276367, "completions/mean_terminated_length": 199.3820358276367, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.03453072585786383, "epoch": 32.16712580348944, "frac_reward_zero_std": 0.19583334326744078, "grad_norm": 0.0283203125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5379511060.0, "reward": 7.938958692550659, "reward_std": 0.37018970847129823, "rewards/combined_reward/mean": 7.938958406448364, "rewards/combined_reward/std": 2.3909681916236876, "step": 52560 }, { "clip_ratio/high_max": 0.0030041319240505498, "clip_ratio/high_mean": 0.001557128268783951, "clip_ratio/low_mean": 0.0009106489168819583, "clip_ratio/low_min": 0.0003248868847246437, "clip_ratio/region_mean": 0.0024677772002178245, "completions/clipped_ratio": 0.0, "completions/max_length": 556.1, "completions/max_terminated_length": 556.1, "completions/mean_length": 184.3664108276367, "completions/mean_terminated_length": 184.3664108276367, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027502129995264114, "epoch": 32.18549127640037, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.004058837890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5382519187.0, "reward": 7.797454214096069, "reward_std": 0.29577395468950274, "rewards/combined_reward/mean": 7.797454357147217, "rewards/combined_reward/std": 2.620434808731079, "step": 52590 }, { "clip_ratio/high_max": 0.002181966714852024, "clip_ratio/high_mean": 0.001263965690304758, "clip_ratio/low_mean": 0.0008729244861266731, "clip_ratio/low_min": 0.00023953764563581595, "clip_ratio/region_mean": 0.0021368901691554737, "completions/clipped_ratio": 0.0, "completions/max_length": 577.7, "completions/max_terminated_length": 577.7, "completions/mean_length": 204.52552947998046, "completions/mean_terminated_length": 204.52552947998046, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.027487543721993763, "epoch": 32.20385674931129, "frac_reward_zero_std": 0.19166667200624943, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5385604725.0, "reward": 8.077548456192016, "reward_std": 0.33687124103307725, "rewards/combined_reward/mean": 8.07754807472229, "rewards/combined_reward/std": 2.358627808094025, "step": 52620 }, { "clip_ratio/high_max": 0.0025323891211883167, "clip_ratio/high_mean": 0.0014315247150079812, "clip_ratio/low_mean": 0.0008177949124122582, "clip_ratio/low_min": 0.00018146008078474553, "clip_ratio/region_mean": 0.0022493196175976966, "completions/clipped_ratio": 0.0, "completions/max_length": 527.4, "completions/max_terminated_length": 527.4, "completions/mean_length": 204.00364990234374, "completions/mean_terminated_length": 204.00364990234374, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.026834237148674826, "epoch": 32.22222222222222, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.01422119140625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5388688259.0, "reward": 8.040960931777954, "reward_std": 0.303023624420166, "rewards/combined_reward/mean": 8.040960597991944, "rewards/combined_reward/std": 2.3196094274520873, "step": 52650 }, { "clip_ratio/high_max": 0.0022117922499698276, "clip_ratio/high_mean": 0.001257969163149634, "clip_ratio/low_mean": 0.0009821929968893528, "clip_ratio/low_min": 0.0002628529849365198, "clip_ratio/region_mean": 0.002240162171074189, "completions/clipped_ratio": 0.0, "completions/max_length": 562.9, "completions/max_terminated_length": 562.9, "completions/mean_length": 202.11562805175782, "completions/mean_terminated_length": 202.11562805175782, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02996875052340329, "epoch": 32.24058769513315, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0244140625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5391764543.0, "reward": 8.101924276351928, "reward_std": 0.35601340755820277, "rewards/combined_reward/mean": 8.101924085617066, "rewards/combined_reward/std": 2.189132606983185, "step": 52680 }, { "clip_ratio/high_max": 0.0024404711235547437, "clip_ratio/high_mean": 0.0014230332439183258, "clip_ratio/low_mean": 0.0010511147113599387, "clip_ratio/low_min": 0.0002936069533461705, "clip_ratio/region_mean": 0.0024741479792282917, "completions/clipped_ratio": 0.0, "completions/max_length": 640.3, "completions/max_terminated_length": 640.3, "completions/mean_length": 196.18177642822266, "completions/mean_terminated_length": 196.18177642822266, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.02867459213982026, "epoch": 32.25895316804408, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.028076171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5394818041.0, "reward": 8.152302598953247, "reward_std": 0.2920657992362976, "rewards/combined_reward/mean": 8.152302360534668, "rewards/combined_reward/std": 2.1248151540756224, "step": 52710 }, { "clip_ratio/high_max": 0.0021571907551939753, "clip_ratio/high_mean": 0.0012761863634295878, "clip_ratio/low_mean": 0.0007769436415401287, "clip_ratio/low_min": 0.00021002063561657754, "clip_ratio/region_mean": 0.0020531300169144137, "completions/clipped_ratio": 0.0, "completions/max_length": 621.8, "completions/max_terminated_length": 621.8, "completions/mean_length": 204.1432357788086, "completions/mean_terminated_length": 204.1432357788086, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.026137834539016087, "epoch": 32.277318640955, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.0120849609375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5397902111.0, "reward": 8.302621746063233, "reward_std": 0.3363124176859856, "rewards/combined_reward/mean": 8.302621269226075, "rewards/combined_reward/std": 1.9974258065223693, "step": 52740 }, { "clip_ratio/high_max": 0.0024719542741271044, "clip_ratio/high_mean": 0.0015425299146348455, "clip_ratio/low_mean": 0.0007437312549276006, "clip_ratio/low_min": 0.00012670901560341008, "clip_ratio/region_mean": 0.0022862611566476215, "completions/clipped_ratio": 0.0, "completions/max_length": 566.3, "completions/max_terminated_length": 566.3, "completions/mean_length": 197.84453887939452, "completions/mean_terminated_length": 197.84453887939452, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.026428337728915115, "epoch": 32.29568411386593, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.01483154296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5400961994.0, "reward": 8.299791860580445, "reward_std": 0.3759906843304634, "rewards/combined_reward/mean": 8.29979157447815, "rewards/combined_reward/std": 1.9578230500221252, "step": 52770 }, { "clip_ratio/high_max": 0.002172192238504067, "clip_ratio/high_mean": 0.001273479263909394, "clip_ratio/low_mean": 0.0009042135537432235, "clip_ratio/low_min": 0.00025265604660186606, "clip_ratio/region_mean": 0.002177692819289708, "completions/clipped_ratio": 0.0, "completions/max_length": 641.7, "completions/max_terminated_length": 641.7, "completions/mean_length": 199.9132873535156, "completions/mean_terminated_length": 199.9132873535156, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.027160867862403393, "epoch": 32.31404958677686, "frac_reward_zero_std": 0.1958333358168602, "grad_norm": 0.00616455078125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5404029821.0, "reward": 8.100742101669312, "reward_std": 0.3086512565612793, "rewards/combined_reward/mean": 8.100741815567016, "rewards/combined_reward/std": 2.114921236038208, "step": 52800 }, { "clip_ratio/high_max": 0.0026240959962403093, "clip_ratio/high_mean": 0.0014126502569221581, "clip_ratio/low_mean": 0.0008157208290261527, "clip_ratio/low_min": 0.00023602622095495463, "clip_ratio/region_mean": 0.00222837109467946, "completions/clipped_ratio": 0.0, "completions/max_length": 519.4, "completions/max_terminated_length": 519.4, "completions/mean_length": 203.9445358276367, "completions/mean_terminated_length": 203.9445358276367, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.03027578432423373, "epoch": 32.33241505968779, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5407113128.0, "reward": 7.999435472488403, "reward_std": 0.2974151447415352, "rewards/combined_reward/mean": 7.999435329437256, "rewards/combined_reward/std": 2.491385757923126, "step": 52830 }, { "clip_ratio/high_max": 0.002395548822823912, "clip_ratio/high_mean": 0.0013215331695391797, "clip_ratio/low_mean": 0.000937713761231862, "clip_ratio/low_min": 0.0002925460072826051, "clip_ratio/region_mean": 0.002259246936106744, "completions/clipped_ratio": 0.0, "completions/max_length": 650.5, "completions/max_terminated_length": 650.5, "completions/mean_length": 198.46797637939454, "completions/mean_terminated_length": 198.46797637939454, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.027004517413054905, "epoch": 32.35078053259871, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5410175405.0, "reward": 7.989316415786743, "reward_std": 0.270861391723156, "rewards/combined_reward/mean": 7.989316129684449, "rewards/combined_reward/std": 2.3428825616836546, "step": 52860 }, { "clip_ratio/high_max": 0.0024702912152861243, "clip_ratio/high_mean": 0.0014483881801425013, "clip_ratio/low_mean": 0.0011713242668823416, "clip_ratio/low_min": 0.0003586070997698698, "clip_ratio/region_mean": 0.0026197124398701513, "completions/clipped_ratio": 0.0, "completions/max_length": 678.6, "completions/max_terminated_length": 678.6, "completions/mean_length": 205.1385482788086, "completions/mean_terminated_length": 205.1385482788086, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.03165665068663657, "epoch": 32.36914600550964, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5413263297.0, "reward": 8.322553014755249, "reward_std": 0.2860825411975384, "rewards/combined_reward/mean": 8.322552728652955, "rewards/combined_reward/std": 1.9438279271125793, "step": 52890 }, { "clip_ratio/high_max": 0.001839583378750831, "clip_ratio/high_mean": 0.0010730951312628652, "clip_ratio/low_mean": 0.0008457363164779963, "clip_ratio/low_min": 0.00020630989286776942, "clip_ratio/region_mean": 0.0019188314501661808, "completions/clipped_ratio": 0.0, "completions/max_length": 547.3, "completions/max_terminated_length": 547.3, "completions/mean_length": 203.2023483276367, "completions/mean_terminated_length": 203.2023483276367, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.024917905335314572, "epoch": 32.38751147842057, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5416343754.0, "reward": 8.19685411453247, "reward_std": 0.2639253243803978, "rewards/combined_reward/mean": 8.196853828430175, "rewards/combined_reward/std": 2.011949437856674, "step": 52920 }, { "clip_ratio/high_max": 0.002775058046487781, "clip_ratio/high_mean": 0.001542472933700386, "clip_ratio/low_mean": 0.001031102138646626, "clip_ratio/low_min": 0.00035666107117625264, "clip_ratio/region_mean": 0.0025735750629489, "completions/clipped_ratio": 0.0, "completions/max_length": 675.9, "completions/max_terminated_length": 675.9, "completions/mean_length": 206.05547180175782, "completions/mean_terminated_length": 206.05547180175782, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.030890358379110694, "epoch": 32.4058769513315, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0198974609375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5419435167.0, "reward": 7.949713039398193, "reward_std": 0.3745655655860901, "rewards/combined_reward/mean": 7.949712896347046, "rewards/combined_reward/std": 2.40702543258667, "step": 52950 }, { "clip_ratio/high_max": 0.002312435112738361, "clip_ratio/high_mean": 0.0013304420608619694, "clip_ratio/low_mean": 0.0008539754333469318, "clip_ratio/low_min": 0.0002745922121296947, "clip_ratio/region_mean": 0.002184417495542827, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 203.33802642822266, "completions/mean_terminated_length": 203.33802642822266, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.02653405931778252, "epoch": 32.42424242424242, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5422516145.0, "reward": 7.932988214492798, "reward_std": 0.35362904369831083, "rewards/combined_reward/mean": 7.932988071441651, "rewards/combined_reward/std": 2.4368468284606934, "step": 52980 }, { "clip_ratio/high_max": 0.0026430421334225685, "clip_ratio/high_mean": 0.0014099278982030229, "clip_ratio/low_mean": 0.0008558633352852969, "clip_ratio/low_min": 0.00016594749128368374, "clip_ratio/region_mean": 0.0022657912336095857, "completions/clipped_ratio": 0.0, "completions/max_length": 487.3, "completions/max_terminated_length": 487.3, "completions/mean_length": 184.18932647705077, "completions/mean_terminated_length": 184.18932647705077, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02567379791289568, "epoch": 32.44260789715335, "frac_reward_zero_std": 0.20833333656191827, "grad_norm": 0.00628662109375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5425523592.0, "reward": 8.31209201812744, "reward_std": 0.26323355808854104, "rewards/combined_reward/mean": 8.312091779708862, "rewards/combined_reward/std": 2.046023726463318, "step": 53010 }, { "clip_ratio/high_max": 0.002154995267240641, "clip_ratio/high_mean": 0.001313783696483976, "clip_ratio/low_mean": 0.001058829865481433, "clip_ratio/low_min": 0.0003430200447231376, "clip_ratio/region_mean": 0.002372613568149973, "completions/clipped_ratio": 0.0, "completions/max_length": 626.1, "completions/max_terminated_length": 626.1, "completions/mean_length": 206.76276702880858, "completions/mean_terminated_length": 206.76276702880858, "completions/min_length": 47.7, "completions/min_terminated_length": 47.7, "entropy": 0.028968638569737474, "epoch": 32.46097337006428, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.0164794921875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5428617721.0, "reward": 8.108874225616455, "reward_std": 0.3462653622031212, "rewards/combined_reward/mean": 8.10887393951416, "rewards/combined_reward/std": 2.3711265444755556, "step": 53040 }, { "clip_ratio/high_max": 0.002202551269632143, "clip_ratio/high_mean": 0.001312475640103609, "clip_ratio/low_mean": 0.0009387458353254867, "clip_ratio/low_min": 0.0002460787114008175, "clip_ratio/region_mean": 0.002251221474822766, "completions/clipped_ratio": 0.0, "completions/max_length": 558.2, "completions/max_terminated_length": 558.2, "completions/mean_length": 201.75339050292968, "completions/mean_terminated_length": 201.75339050292968, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02725484052207321, "epoch": 32.47933884297521, "frac_reward_zero_std": 0.19166667126119136, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5431692614.0, "reward": 8.07805585861206, "reward_std": 0.2736371695995331, "rewards/combined_reward/mean": 8.07805576324463, "rewards/combined_reward/std": 2.154739201068878, "step": 53070 }, { "clip_ratio/high_max": 0.002564543489764522, "clip_ratio/high_mean": 0.0014165784767101286, "clip_ratio/low_mean": 0.0009967049684443433, "clip_ratio/low_min": 0.000318936223872394, "clip_ratio/region_mean": 0.002413283459706387, "completions/clipped_ratio": 0.0, "completions/max_length": 529.3, "completions/max_terminated_length": 529.3, "completions/mean_length": 202.77734985351563, "completions/mean_terminated_length": 202.77734985351563, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02905179113925745, "epoch": 32.49770431588613, "frac_reward_zero_std": 0.2083333373069763, "grad_norm": 0.003997802734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5434771439.0, "reward": 8.037032318115234, "reward_std": 0.2842897571623325, "rewards/combined_reward/mean": 8.03703203201294, "rewards/combined_reward/std": 2.3686881184577944, "step": 53100 }, { "clip_ratio/high_max": 0.0021721652510071484, "clip_ratio/high_mean": 0.0012379410696060707, "clip_ratio/low_mean": 0.0009360725114675005, "clip_ratio/low_min": 0.0003177992989852404, "clip_ratio/region_mean": 0.0021740135896834547, "completions/clipped_ratio": 0.0, "completions/max_length": 588.6, "completions/max_terminated_length": 588.6, "completions/mean_length": 196.97500610351562, "completions/mean_terminated_length": 196.97500610351562, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.024836528409893314, "epoch": 32.51606978879706, "frac_reward_zero_std": 0.1916666716337204, "grad_norm": 0.01470947265625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5437827983.0, "reward": 8.123004150390624, "reward_std": 0.2983556821942329, "rewards/combined_reward/mean": 8.123003816604614, "rewards/combined_reward/std": 2.141700768470764, "step": 53130 }, { "clip_ratio/high_max": 0.0025499389094572203, "clip_ratio/high_mean": 0.0015136039941959704, "clip_ratio/low_mean": 0.0009166540126898326, "clip_ratio/low_min": 0.00031246520860198264, "clip_ratio/region_mean": 0.0024302580112513777, "completions/clipped_ratio": 0.0, "completions/max_length": 632.2, "completions/max_terminated_length": 632.2, "completions/mean_length": 212.56849517822266, "completions/mean_terminated_length": 212.56849517822266, "completions/min_length": 43.3, "completions/min_terminated_length": 43.3, "entropy": 0.030845538852736352, "epoch": 32.53443526170799, "frac_reward_zero_std": 0.16666667349636555, "grad_norm": 0.004547119140625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5440944406.0, "reward": 8.062848663330078, "reward_std": 0.29813436791300774, "rewards/combined_reward/mean": 8.062848329544067, "rewards/combined_reward/std": 2.233007049560547, "step": 53160 }, { "clip_ratio/high_max": 0.0026188284156281345, "clip_ratio/high_mean": 0.0013642124832889144, "clip_ratio/low_mean": 0.0010054513390059584, "clip_ratio/low_min": 0.00029279873682147204, "clip_ratio/region_mean": 0.002369663823628798, "completions/clipped_ratio": 0.0, "completions/max_length": 597.1, "completions/max_terminated_length": 597.1, "completions/mean_length": 196.25443267822266, "completions/mean_terminated_length": 196.25443267822266, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.029435442682976525, "epoch": 32.55280073461892, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.005218505859375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5443998183.0, "reward": 8.045079708099365, "reward_std": 0.29573310166597366, "rewards/combined_reward/mean": 8.04507942199707, "rewards/combined_reward/std": 2.461572599411011, "step": 53190 }, { "clip_ratio/high_max": 0.0023334904030586283, "clip_ratio/high_mean": 0.0013586905998333046, "clip_ratio/low_mean": 0.0009632634952140507, "clip_ratio/low_min": 0.0002781145582654669, "clip_ratio/region_mean": 0.0023219540647308654, "completions/clipped_ratio": 0.0, "completions/max_length": 521.4, "completions/max_terminated_length": 521.4, "completions/mean_length": 191.80989990234374, "completions/mean_terminated_length": 191.80989990234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.027770387521013618, "epoch": 32.57116620752984, "frac_reward_zero_std": 0.20416667684912682, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5447034893.0, "reward": 8.026461267471314, "reward_std": 0.3249573685228825, "rewards/combined_reward/mean": 8.026460886001587, "rewards/combined_reward/std": 2.284532403945923, "step": 53220 }, { "clip_ratio/high_max": 0.0025815020208635056, "clip_ratio/high_mean": 0.0014218277972152767, "clip_ratio/low_mean": 0.0009959136505737357, "clip_ratio/low_min": 0.00034182636494127413, "clip_ratio/region_mean": 0.0024177414617345978, "completions/clipped_ratio": 0.0, "completions/max_length": 590.5, "completions/max_terminated_length": 590.5, "completions/mean_length": 192.74662017822266, "completions/mean_terminated_length": 192.74662017822266, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.03022084357216954, "epoch": 32.58953168044077, "frac_reward_zero_std": 0.12916666865348816, "grad_norm": 0.0240478515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5450075200.0, "reward": 7.957585096359253, "reward_std": 0.28211243003606795, "rewards/combined_reward/mean": 7.957585048675537, "rewards/combined_reward/std": 2.521455669403076, "step": 53250 }, { "clip_ratio/high_max": 0.0026246949719885984, "clip_ratio/high_mean": 0.001430657880215828, "clip_ratio/low_mean": 0.001128272200791495, "clip_ratio/low_min": 0.00031808175820818483, "clip_ratio/region_mean": 0.002558930078521371, "completions/clipped_ratio": 0.0, "completions/max_length": 575.5, "completions/max_terminated_length": 575.5, "completions/mean_length": 203.73567962646484, "completions/mean_terminated_length": 203.73567962646484, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.031132113297159474, "epoch": 32.6078971533517, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5453157705.0, "reward": 7.990364646911621, "reward_std": 0.39480437636375426, "rewards/combined_reward/mean": 7.990364360809326, "rewards/combined_reward/std": 2.2147345304489137, "step": 53280 }, { "clip_ratio/high_max": 0.00264683477289509, "clip_ratio/high_mean": 0.0014775245758452608, "clip_ratio/low_mean": 0.0011106496443971992, "clip_ratio/low_min": 0.00042386460869844693, "clip_ratio/region_mean": 0.002588174195746736, "completions/clipped_ratio": 0.0, "completions/max_length": 557.7, "completions/max_terminated_length": 557.7, "completions/mean_length": 191.32760925292968, "completions/mean_terminated_length": 191.32760925292968, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.03270669863559306, "epoch": 32.62626262626263, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5456192563.0, "reward": 8.057622909545898, "reward_std": 0.35781753808259964, "rewards/combined_reward/mean": 8.057622814178467, "rewards/combined_reward/std": 2.065883696079254, "step": 53310 }, { "clip_ratio/high_max": 0.002480556829929507, "clip_ratio/high_mean": 0.001465316942873566, "clip_ratio/low_mean": 0.0011223645222344202, "clip_ratio/low_min": 0.0003286779256692777, "clip_ratio/region_mean": 0.002587681475658125, "completions/clipped_ratio": 0.0, "completions/max_length": 560.1, "completions/max_terminated_length": 560.1, "completions/mean_length": 204.0408905029297, "completions/mean_terminated_length": 204.0408905029297, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.0316243610655268, "epoch": 32.64462809917355, "frac_reward_zero_std": 0.1708333358168602, "grad_norm": 0.017333984375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5459276240.0, "reward": 8.146624231338501, "reward_std": 0.32892731577157974, "rewards/combined_reward/mean": 8.146623945236206, "rewards/combined_reward/std": 2.227451169490814, "step": 53340 }, { "clip_ratio/high_max": 0.0030183081231371033, "clip_ratio/high_mean": 0.0017525982394242115, "clip_ratio/low_mean": 0.0010801315396141338, "clip_ratio/low_min": 0.00023328537766550048, "clip_ratio/region_mean": 0.0028327297787958134, "completions/clipped_ratio": 0.0, "completions/max_length": 593.7, "completions/max_terminated_length": 593.7, "completions/mean_length": 198.48099365234376, "completions/mean_terminated_length": 198.48099365234376, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03403111433920761, "epoch": 32.66299357208448, "frac_reward_zero_std": 0.17500000670552254, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5462338567.0, "reward": 7.810449075698853, "reward_std": 0.4017158403992653, "rewards/combined_reward/mean": 7.810448741912841, "rewards/combined_reward/std": 2.3873268246650694, "step": 53370 }, { "clip_ratio/high_max": 0.0022890221422130707, "clip_ratio/high_mean": 0.001312600157931835, "clip_ratio/low_mean": 0.0010210232911049388, "clip_ratio/low_min": 0.0003450155711713402, "clip_ratio/region_mean": 0.002333623458495519, "completions/clipped_ratio": 0.0, "completions/max_length": 532.4, "completions/max_terminated_length": 532.4, "completions/mean_length": 201.3455780029297, "completions/mean_terminated_length": 201.3455780029297, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.029069084270546833, "epoch": 32.68135904499541, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.0172119140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5465411894.0, "reward": 8.005759143829346, "reward_std": 0.30189111530780793, "rewards/combined_reward/mean": 8.005758810043336, "rewards/combined_reward/std": 2.30200252532959, "step": 53400 }, { "clip_ratio/high_max": 0.0022937989308654022, "clip_ratio/high_mean": 0.0012813965101183082, "clip_ratio/low_mean": 0.0009064896636118647, "clip_ratio/low_min": 0.00021286234380871367, "clip_ratio/region_mean": 0.002187886173730173, "completions/clipped_ratio": 0.0, "completions/max_length": 510.7, "completions/max_terminated_length": 510.7, "completions/mean_length": 183.69635772705078, "completions/mean_terminated_length": 183.69635772705078, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02731831297278404, "epoch": 32.69972451790633, "frac_reward_zero_std": 0.19166667088866235, "grad_norm": 0.00860595703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5468417448.0, "reward": 8.118201732635498, "reward_std": 0.30887471735477445, "rewards/combined_reward/mean": 8.118201303482056, "rewards/combined_reward/std": 2.2361913323402405, "step": 53430 }, { "clip_ratio/high_max": 0.0021363309851343123, "clip_ratio/high_mean": 0.0012323839342570863, "clip_ratio/low_mean": 0.0011457943549127473, "clip_ratio/low_min": 0.00038446934365007717, "clip_ratio/region_mean": 0.0023781782981435145, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 211.9494842529297, "completions/mean_terminated_length": 211.9494842529297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030202153076728185, "epoch": 32.71808999081726, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.0115966796875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5471531494.0, "reward": 8.032287788391113, "reward_std": 0.2966591238975525, "rewards/combined_reward/mean": 8.03228759765625, "rewards/combined_reward/std": 2.3654632210731505, "step": 53460 }, { "clip_ratio/high_max": 0.002234366810929108, "clip_ratio/high_mean": 0.0013167821388681963, "clip_ratio/low_mean": 0.0010214288886345457, "clip_ratio/low_min": 0.00020609533045596134, "clip_ratio/region_mean": 0.0023382110249561567, "completions/clipped_ratio": 0.0, "completions/max_length": 566.2, "completions/max_terminated_length": 566.2, "completions/mean_length": 198.72604522705078, "completions/mean_terminated_length": 198.72604522705078, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.030580693297088145, "epoch": 32.73645546372819, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.00830078125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5474594762.0, "reward": 7.863268136978149, "reward_std": 0.3699320197105408, "rewards/combined_reward/mean": 7.8632677555084225, "rewards/combined_reward/std": 2.313759219646454, "step": 53490 }, { "clip_ratio/high_max": 0.003359530774469022, "clip_ratio/high_mean": 0.0016411720047472046, "clip_ratio/low_mean": 0.0010771369040109373, "clip_ratio/low_min": 0.00024037333350861444, "clip_ratio/region_mean": 0.0027183089123961205, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 204.9507858276367, "completions/mean_terminated_length": 204.9507858276367, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03333160847735902, "epoch": 32.75482093663912, "frac_reward_zero_std": 0.1833333384245634, "grad_norm": 0.009765625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5477681933.0, "reward": 7.875512027740479, "reward_std": 0.29400091916322707, "rewards/combined_reward/mean": 7.8755117893219, "rewards/combined_reward/std": 2.4159158170223236, "step": 53520 }, { "clip_ratio/high_max": 0.0023077543461113236, "clip_ratio/high_mean": 0.0013690209848088368, "clip_ratio/low_mean": 0.0008748712128484234, "clip_ratio/low_min": 0.00019525375561594652, "clip_ratio/region_mean": 0.002243892192685356, "completions/clipped_ratio": 0.0, "completions/max_length": 672.5, "completions/max_terminated_length": 672.5, "completions/mean_length": 195.54140930175782, "completions/mean_terminated_length": 195.54140930175782, "completions/min_length": 41.9, "completions/min_terminated_length": 41.9, "entropy": 0.028548394748941063, "epoch": 32.77318640955004, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.004302978515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5480732972.0, "reward": 8.036963605880738, "reward_std": 0.26523547917604445, "rewards/combined_reward/mean": 8.036963415145873, "rewards/combined_reward/std": 2.355649596452713, "step": 53550 }, { "clip_ratio/high_max": 0.0025134471252385993, "clip_ratio/high_mean": 0.0014092848003807981, "clip_ratio/low_mean": 0.0009337717817591814, "clip_ratio/low_min": 0.00033787607098929585, "clip_ratio/region_mean": 0.0023430565745608573, "completions/clipped_ratio": 0.0, "completions/max_length": 572.6, "completions/max_terminated_length": 572.6, "completions/mean_length": 205.79141082763672, "completions/mean_terminated_length": 205.79141082763672, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.027612863294780254, "epoch": 32.79155188246097, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.012939453125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5483823371.0, "reward": 7.972214317321777, "reward_std": 0.3930189684033394, "rewards/combined_reward/mean": 7.972214126586914, "rewards/combined_reward/std": 2.29901043176651, "step": 53580 }, { "clip_ratio/high_max": 0.0026715405052527784, "clip_ratio/high_mean": 0.0015437512680364306, "clip_ratio/low_mean": 0.0010090622911472261, "clip_ratio/low_min": 0.00025945926827262154, "clip_ratio/region_mean": 0.0025528135748269657, "completions/clipped_ratio": 0.0, "completions/max_length": 595.1, "completions/max_terminated_length": 595.1, "completions/mean_length": 200.1554733276367, "completions/mean_terminated_length": 200.1554733276367, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03159809980231027, "epoch": 32.8099173553719, "frac_reward_zero_std": 0.12083333469927311, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5486892128.0, "reward": 8.205611085891723, "reward_std": 0.3923216238617897, "rewards/combined_reward/mean": 8.20561079978943, "rewards/combined_reward/std": 1.957983523607254, "step": 53610 }, { "clip_ratio/high_max": 0.0020838262023365433, "clip_ratio/high_mean": 0.0011796997966788088, "clip_ratio/low_mean": 0.0010441436426238701, "clip_ratio/low_min": 0.00036495267922873607, "clip_ratio/region_mean": 0.002223843439423945, "completions/clipped_ratio": 0.0, "completions/max_length": 549.7, "completions/max_terminated_length": 549.7, "completions/mean_length": 211.59271545410155, "completions/mean_terminated_length": 211.59271545410155, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.027955582443003852, "epoch": 32.82828282828283, "frac_reward_zero_std": 0.14166666977107525, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 5490004804.0, "reward": 7.708446645736695, "reward_std": 0.37281434834003446, "rewards/combined_reward/mean": 7.708446502685547, "rewards/combined_reward/std": 2.5461625456809998, "step": 53640 }, { "clip_ratio/high_max": 0.0023937921789183745, "clip_ratio/high_mean": 0.0012886135996571586, "clip_ratio/low_mean": 0.0007735766311573873, "clip_ratio/low_min": 0.00027121878520119935, "clip_ratio/region_mean": 0.0020621902212345352, "completions/clipped_ratio": 0.0, "completions/max_length": 682.3, "completions/max_terminated_length": 682.3, "completions/mean_length": 205.1304733276367, "completions/mean_terminated_length": 205.1304733276367, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.025848139155035216, "epoch": 32.84664830119375, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.0281982421875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5493092665.0, "reward": 7.829952621459961, "reward_std": 0.3476113364100456, "rewards/combined_reward/mean": 7.8299524784088135, "rewards/combined_reward/std": 2.423458981513977, "step": 53670 }, { "clip_ratio/high_max": 0.0018811653569476524, "clip_ratio/high_mean": 0.0010641718445792018, "clip_ratio/low_mean": 0.0007609582414564405, "clip_ratio/low_min": 0.00019336546465638095, "clip_ratio/region_mean": 0.00182513008415602, "completions/clipped_ratio": 0.0, "completions/max_length": 639.5, "completions/max_terminated_length": 639.5, "completions/mean_length": 195.47188262939454, "completions/mean_terminated_length": 195.47188262939454, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.025721868119823434, "epoch": 32.86501377410468, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.015380859375, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 5496143437.0, "reward": 7.882122611999511, "reward_std": 0.2652942694723606, "rewards/combined_reward/mean": 7.882122421264649, "rewards/combined_reward/std": 2.5180380582809447, "step": 53700 }, { "clip_ratio/high_max": 0.002411211062523459, "clip_ratio/high_mean": 0.0013629622980564212, "clip_ratio/low_mean": 0.0009436362563671234, "clip_ratio/low_min": 0.0002625582254646967, "clip_ratio/region_mean": 0.002306598565580013, "completions/clipped_ratio": 0.0, "completions/max_length": 643.3, "completions/max_terminated_length": 643.3, "completions/mean_length": 208.2565170288086, "completions/mean_terminated_length": 208.2565170288086, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.029260248007873694, "epoch": 32.88337924701561, "frac_reward_zero_std": 0.1500000014901161, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.0025, "num_tokens": 5499243302.0, "reward": 8.15282220840454, "reward_std": 0.3762630209326744, "rewards/combined_reward/mean": 8.152821969985961, "rewards/combined_reward/std": 2.2326192617416383, "step": 53730 }, { "clip_ratio/high_max": 0.002544823905433683, "clip_ratio/high_mean": 0.0014871781541539045, "clip_ratio/low_mean": 0.00105722918669926, "clip_ratio/low_min": 0.0003655462069824959, "clip_ratio/region_mean": 0.0025444073471589945, "completions/clipped_ratio": 0.0, "completions/max_length": 625.7, "completions/max_terminated_length": 625.7, "completions/mean_length": 211.6682357788086, "completions/mean_terminated_length": 211.6682357788086, "completions/min_length": 50.4, "completions/min_terminated_length": 50.4, "entropy": 0.03147997343136619, "epoch": 32.90174471992654, "frac_reward_zero_std": 0.12083333618938923, "grad_norm": 0.031494140625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5502356268.0, "reward": 8.04659628868103, "reward_std": 0.4425238460302353, "rewards/combined_reward/mean": 8.046596193313599, "rewards/combined_reward/std": 2.2724488019943236, "step": 53760 }, { "clip_ratio/high_max": 0.002132866318667463, "clip_ratio/high_mean": 0.0012357050460802081, "clip_ratio/low_mean": 0.000865641501028828, "clip_ratio/low_min": 0.00017937210432137363, "clip_ratio/region_mean": 0.0021013465375290252, "completions/clipped_ratio": 0.0, "completions/max_length": 560.5, "completions/max_terminated_length": 560.5, "completions/mean_length": 210.2007858276367, "completions/mean_terminated_length": 210.2007858276367, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.026912980573251845, "epoch": 32.92011019283746, "frac_reward_zero_std": 0.17500000521540643, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5505463599.0, "reward": 7.971912622451782, "reward_std": 0.35111969858407976, "rewards/combined_reward/mean": 7.971912860870361, "rewards/combined_reward/std": 2.3057084321975707, "step": 53790 }, { "clip_ratio/high_max": 0.0024311571428067206, "clip_ratio/high_mean": 0.0014237153163776384, "clip_ratio/low_mean": 0.0008627868555777241, "clip_ratio/low_min": 0.00023175630849436857, "clip_ratio/region_mean": 0.0022865021683780165, "completions/clipped_ratio": 0.0, "completions/max_length": 570.1, "completions/max_terminated_length": 570.1, "completions/mean_length": 205.05026397705078, "completions/mean_terminated_length": 205.05026397705078, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 0.02797905116652449, "epoch": 32.93847566574839, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5508551152.0, "reward": 8.118936395645141, "reward_std": 0.327293398976326, "rewards/combined_reward/mean": 8.118936204910279, "rewards/combined_reward/std": 2.3272502422332764, "step": 53820 }, { "clip_ratio/high_max": 0.0026618468575179575, "clip_ratio/high_mean": 0.0015226592874872343, "clip_ratio/low_mean": 0.0010847001747606555, "clip_ratio/low_min": 0.0003024277541650614, "clip_ratio/region_mean": 0.0026073594674623263, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 210.85156707763673, "completions/mean_terminated_length": 210.85156707763673, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.0316553129038463, "epoch": 32.95684113865932, "frac_reward_zero_std": 0.17916666865348815, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5511660982.0, "reward": 7.947271394729614, "reward_std": 0.3438254237174988, "rewards/combined_reward/mean": 7.947271156311035, "rewards/combined_reward/std": 2.318809449672699, "step": 53850 }, { "clip_ratio/high_max": 0.0023581282075610943, "clip_ratio/high_mean": 0.0013578302937579186, "clip_ratio/low_mean": 0.0009652336695580744, "clip_ratio/low_min": 0.00031091454196333264, "clip_ratio/region_mean": 0.0023230639320293753, "completions/clipped_ratio": 0.0, "completions/max_length": 596.5, "completions/max_terminated_length": 596.5, "completions/mean_length": 205.2760437011719, "completions/mean_terminated_length": 205.2760437011719, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.0295749584135289, "epoch": 32.97520661157025, "frac_reward_zero_std": 0.19166667349636554, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5514749402.0, "reward": 7.929133939743042, "reward_std": 0.35569829046726226, "rewards/combined_reward/mean": 7.929133796691895, "rewards/combined_reward/std": 2.4556315898895265, "step": 53880 }, { "clip_ratio/high_max": 0.0024146074889965044, "clip_ratio/high_mean": 0.001334411045051335, "clip_ratio/low_mean": 0.0008957835678302217, "clip_ratio/low_min": 0.0002220622850776029, "clip_ratio/region_mean": 0.0022301946252506846, "completions/clipped_ratio": 0.0, "completions/max_length": 483.3, "completions/max_terminated_length": 483.3, "completions/mean_length": 189.60130767822267, "completions/mean_terminated_length": 189.60130767822267, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "entropy": 0.02776486035436392, "epoch": 32.99357208448117, "frac_reward_zero_std": 0.18333333656191825, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5517777631.0, "reward": 8.171882724761963, "reward_std": 0.3397939786314964, "rewards/combined_reward/mean": 8.1718825340271, "rewards/combined_reward/std": 2.203820288181305, "step": 53910 }, { "clip_ratio/high_max": 0.0018976673190837126, "clip_ratio/high_mean": 0.0011353159082364843, "clip_ratio/low_mean": 0.0009555605355348296, "clip_ratio/low_min": 0.00027706889633949595, "clip_ratio/region_mean": 0.0020908764593097996, "completions/clipped_ratio": 0.0, "completions/max_length": 566.1, "completions/max_terminated_length": 566.1, "completions/mean_length": 201.79141082763672, "completions/mean_terminated_length": 201.79141082763672, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.03159571077548346, "epoch": 33.01163146617692, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.00982666015625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5520852670.0, "reward": 8.075923442840576, "reward_std": 0.3590974807739258, "rewards/combined_reward/mean": 8.075923252105714, "rewards/combined_reward/std": 2.1504900455474854, "step": 53940 }, { "clip_ratio/high_max": 0.0019750095321796834, "clip_ratio/high_mean": 0.0010065831045115677, "clip_ratio/low_mean": 0.0006701179500547975, "clip_ratio/low_min": 0.00019019419560208916, "clip_ratio/region_mean": 0.0016767010526867428, "completions/clipped_ratio": 0.0, "completions/max_length": 646.3, "completions/max_terminated_length": 646.3, "completions/mean_length": 197.22109832763672, "completions/mean_terminated_length": 197.22109832763672, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.02890113862231374, "epoch": 33.02999693908785, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.017822265625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5523910159.0, "reward": 8.040721225738526, "reward_std": 0.31587970405817034, "rewards/combined_reward/mean": 8.040721082687378, "rewards/combined_reward/std": 2.362937569618225, "step": 53970 }, { "clip_ratio/high_max": 0.0015688248041745587, "clip_ratio/high_mean": 0.0008912118851488533, "clip_ratio/low_mean": 0.0006933733227924676, "clip_ratio/low_min": 0.00019987561851545857, "clip_ratio/region_mean": 0.0015845851977549805, "completions/clipped_ratio": 0.0, "completions/max_length": 590.1, "completions/max_terminated_length": 590.1, "completions/mean_length": 204.1867248535156, "completions/mean_terminated_length": 204.1867248535156, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.028755820589140058, "epoch": 33.048362411998774, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0133056640625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5526994396.0, "reward": 7.83561863899231, "reward_std": 0.3435852140188217, "rewards/combined_reward/mean": 7.835618400573731, "rewards/combined_reward/std": 2.29869704246521, "step": 54000 }, { "clip_ratio/high_max": 0.0019269179019223278, "clip_ratio/high_mean": 0.0010667210842560356, "clip_ratio/low_mean": 0.0007536615269297423, "clip_ratio/low_min": 0.00019556808304817727, "clip_ratio/region_mean": 0.0018203826305883315, "completions/clipped_ratio": 0.0, "completions/max_length": 532.9, "completions/max_terminated_length": 532.9, "completions/mean_length": 183.89687957763672, "completions/mean_terminated_length": 183.89687957763672, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.027315496125568946, "epoch": 33.06611570247934, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.0216064453125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5530000720.0, "reward": 8.164693546295165, "reward_std": 0.28323918879032134, "rewards/combined_reward/mean": 8.16469340324402, "rewards/combined_reward/std": 2.205184745788574, "step": 54030 }, { "clip_ratio/high_max": 0.001671166959082863, "clip_ratio/high_mean": 0.0009805776695429812, "clip_ratio/low_mean": 0.0007426324523597335, "clip_ratio/low_min": 0.000262186075997306, "clip_ratio/region_mean": 0.001723210124085502, "completions/clipped_ratio": 0.0, "completions/max_length": 594.9, "completions/max_terminated_length": 594.9, "completions/mean_length": 201.9507843017578, "completions/mean_terminated_length": 201.9507843017578, "completions/min_length": 38.3, "completions/min_terminated_length": 38.3, "entropy": 0.028699765889905395, "epoch": 33.084481175390266, "frac_reward_zero_std": 0.15833334103226662, "grad_norm": 0.00933837890625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5533076371.0, "reward": 7.886908531188965, "reward_std": 0.32335265427827836, "rewards/combined_reward/mean": 7.886908388137817, "rewards/combined_reward/std": 2.5092972040176393, "step": 54060 }, { "clip_ratio/high_max": 0.0016244885029057818, "clip_ratio/high_mean": 0.0008970078308872568, "clip_ratio/low_mean": 0.000665123061844497, "clip_ratio/low_min": 0.000191680471956109, "clip_ratio/region_mean": 0.0015621308994013816, "completions/clipped_ratio": 0.0, "completions/max_length": 568.9, "completions/max_terminated_length": 568.9, "completions/mean_length": 201.64141082763672, "completions/mean_terminated_length": 201.64141082763672, "completions/min_length": 47.9, "completions/min_terminated_length": 47.9, "entropy": 0.026066891895607115, "epoch": 33.102846648301195, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5536150834.0, "reward": 8.004878091812134, "reward_std": 0.2666877344250679, "rewards/combined_reward/mean": 8.004878044128418, "rewards/combined_reward/std": 2.2134251832962035, "step": 54090 }, { "clip_ratio/high_max": 0.0021975409690639935, "clip_ratio/high_mean": 0.0012487668805988505, "clip_ratio/low_mean": 0.0008036233810950459, "clip_ratio/low_min": 0.00020925083954352884, "clip_ratio/region_mean": 0.002052390269818716, "completions/clipped_ratio": 0.0, "completions/max_length": 714.7, "completions/max_terminated_length": 714.7, "completions/mean_length": 207.05026550292968, "completions/mean_terminated_length": 207.05026550292968, "completions/min_length": 40.3, "completions/min_terminated_length": 40.3, "entropy": 0.03320434124519427, "epoch": 33.121212121212125, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.0284423828125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5539246067.0, "reward": 7.842648839950561, "reward_std": 0.3623768538236618, "rewards/combined_reward/mean": 7.842648696899414, "rewards/combined_reward/std": 2.494632089138031, "step": 54120 }, { "clip_ratio/high_max": 0.002085250632808311, "clip_ratio/high_mean": 0.0011425441846464916, "clip_ratio/low_mean": 0.0007447362222592346, "clip_ratio/low_min": 0.00015774839060516871, "clip_ratio/region_mean": 0.0018872803981745772, "completions/clipped_ratio": 0.0, "completions/max_length": 600.9, "completions/max_terminated_length": 600.9, "completions/mean_length": 197.4726593017578, "completions/mean_terminated_length": 197.4726593017578, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03021583353790144, "epoch": 33.13957759412305, "frac_reward_zero_std": 0.15000000335276126, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5542304522.0, "reward": 8.15880036354065, "reward_std": 0.3377731621265411, "rewards/combined_reward/mean": 8.158800172805787, "rewards/combined_reward/std": 2.15181679725647, "step": 54150 }, { "clip_ratio/high_max": 0.002268156737651831, "clip_ratio/high_mean": 0.0011516301485244185, "clip_ratio/low_mean": 0.0007986833305039908, "clip_ratio/low_min": 0.00018448508708388545, "clip_ratio/region_mean": 0.0019503134763605582, "completions/clipped_ratio": 0.0, "completions/max_length": 635.5, "completions/max_terminated_length": 635.5, "completions/mean_length": 199.81250457763673, "completions/mean_terminated_length": 199.81250457763673, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.028734662507971128, "epoch": 33.157943067033976, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5545371962.0, "reward": 8.042540693283081, "reward_std": 0.36234794557094574, "rewards/combined_reward/mean": 8.042540407180786, "rewards/combined_reward/std": 2.216623032093048, "step": 54180 }, { "clip_ratio/high_max": 0.00202267936062223, "clip_ratio/high_mean": 0.0011638864561973605, "clip_ratio/low_mean": 0.0006931274998351, "clip_ratio/low_min": 0.00019170056815103937, "clip_ratio/region_mean": 0.001857013951424354, "completions/clipped_ratio": 0.0, "completions/max_length": 579.2, "completions/max_terminated_length": 579.2, "completions/mean_length": 193.42969360351563, "completions/mean_terminated_length": 193.42969360351563, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.028363737006050844, "epoch": 33.176308539944905, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5548414892.0, "reward": 8.019516515731812, "reward_std": 0.33573122769594194, "rewards/combined_reward/mean": 8.019516324996948, "rewards/combined_reward/std": 2.4208574414253237, "step": 54210 }, { "clip_ratio/high_max": 0.001760632107713415, "clip_ratio/high_mean": 0.000949834370597576, "clip_ratio/low_mean": 0.0006479435659760686, "clip_ratio/low_min": 0.00013917081499433456, "clip_ratio/region_mean": 0.0015977779349971874, "completions/clipped_ratio": 0.0, "completions/max_length": 605.6, "completions/max_terminated_length": 605.6, "completions/mean_length": 192.89167175292968, "completions/mean_terminated_length": 192.89167175292968, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.026372431366083523, "epoch": 33.19467401285583, "frac_reward_zero_std": 0.20000000707805157, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5551455756.0, "reward": 8.033796501159667, "reward_std": 0.30417789816856383, "rewards/combined_reward/mean": 8.033796167373657, "rewards/combined_reward/std": 2.3280038475990295, "step": 54240 }, { "clip_ratio/high_max": 0.0018983514261587213, "clip_ratio/high_mean": 0.0010682000036467798, "clip_ratio/low_mean": 0.0007719104357723457, "clip_ratio/low_min": 0.00014833206684367422, "clip_ratio/region_mean": 0.0018401104408743171, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 202.5119857788086, "completions/mean_terminated_length": 202.5119857788086, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.02732244028399388, "epoch": 33.21303948576676, "frac_reward_zero_std": 0.2250000074505806, "grad_norm": 0.00927734375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5554533562.0, "reward": 7.947984981536865, "reward_std": 0.30147129148244856, "rewards/combined_reward/mean": 7.9479845523834225, "rewards/combined_reward/std": 2.4705838441848753, "step": 54270 }, { "clip_ratio/high_max": 0.0020650493669866895, "clip_ratio/high_mean": 0.0011854754823919695, "clip_ratio/low_mean": 0.0008037209142154703, "clip_ratio/low_min": 0.0002606632673026373, "clip_ratio/region_mean": 0.0019891964077639085, "completions/clipped_ratio": 0.0, "completions/max_length": 510.7, "completions/max_terminated_length": 510.7, "completions/mean_length": 198.2987045288086, "completions/mean_terminated_length": 198.2987045288086, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.028957806869099538, "epoch": 33.231404958677686, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.00762939453125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5557595189.0, "reward": 8.154308748245239, "reward_std": 0.3724937975406647, "rewards/combined_reward/mean": 8.154308366775513, "rewards/combined_reward/std": 2.0422839164733886, "step": 54300 }, { "clip_ratio/high_max": 0.0018687927872330572, "clip_ratio/high_mean": 0.0010095580168126617, "clip_ratio/low_mean": 0.0007380236268848724, "clip_ratio/low_min": 0.00019354663066527186, "clip_ratio/region_mean": 0.001747581625143842, "completions/clipped_ratio": 0.0, "completions/max_length": 578.9, "completions/max_terminated_length": 578.9, "completions/mean_length": 191.2643295288086, "completions/mean_terminated_length": 191.2643295288086, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.026949991782506306, "epoch": 33.249770431588615, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.00970458984375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5560629804.0, "reward": 8.125395441055298, "reward_std": 0.271054907143116, "rewards/combined_reward/mean": 8.125394916534423, "rewards/combined_reward/std": 2.155404198169708, "step": 54330 }, { "clip_ratio/high_max": 0.002070440136594698, "clip_ratio/high_mean": 0.0011140177945587009, "clip_ratio/low_mean": 0.0008281750087917317, "clip_ratio/low_min": 0.00022934480681821394, "clip_ratio/region_mean": 0.0019421928074734752, "completions/clipped_ratio": 0.0, "completions/max_length": 554.9, "completions/max_terminated_length": 554.9, "completions/mean_length": 207.0737045288086, "completions/mean_terminated_length": 207.0737045288086, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03064962485805154, "epoch": 33.26813590449954, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.018310546875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5563725127.0, "reward": 8.135510396957397, "reward_std": 0.3336356818675995, "rewards/combined_reward/mean": 8.135510349273682, "rewards/combined_reward/std": 2.116107499599457, "step": 54360 }, { "clip_ratio/high_max": 0.0020344138780880408, "clip_ratio/high_mean": 0.0011518948762386572, "clip_ratio/low_mean": 0.0007215045113601567, "clip_ratio/low_min": 0.00020658264790351192, "clip_ratio/region_mean": 0.00187339938711375, "completions/clipped_ratio": 0.0, "completions/max_length": 634.2, "completions/max_terminated_length": 634.2, "completions/mean_length": 197.04193267822265, "completions/mean_terminated_length": 197.04193267822265, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.028491072884450356, "epoch": 33.28650137741047, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.00946044921875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5566781928.0, "reward": 7.9497757911682125, "reward_std": 0.3011281430721283, "rewards/combined_reward/mean": 7.949775648117066, "rewards/combined_reward/std": 2.315136432647705, "step": 54390 }, { "clip_ratio/high_max": 0.0018545303081433909, "clip_ratio/high_mean": 0.0011170551713197104, "clip_ratio/low_mean": 0.000750665990926791, "clip_ratio/low_min": 0.00024857510021926524, "clip_ratio/region_mean": 0.0018677211560619374, "completions/clipped_ratio": 0.0, "completions/max_length": 539.9, "completions/max_terminated_length": 539.9, "completions/mean_length": 214.3244842529297, "completions/mean_terminated_length": 214.3244842529297, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.032079862477257845, "epoch": 33.304866850321396, "frac_reward_zero_std": 0.15416667088866234, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5569905094.0, "reward": 8.098084974288941, "reward_std": 0.3333271235227585, "rewards/combined_reward/mean": 8.098084974288941, "rewards/combined_reward/std": 2.1034961462020876, "step": 54420 }, { "clip_ratio/high_max": 0.0018613284725385408, "clip_ratio/high_mean": 0.0010400119843931557, "clip_ratio/low_mean": 0.0007269776483250705, "clip_ratio/low_min": 0.00020797512794767196, "clip_ratio/region_mean": 0.0017669896362349391, "completions/clipped_ratio": 0.0, "completions/max_length": 619.4, "completions/max_terminated_length": 619.4, "completions/mean_length": 203.0773483276367, "completions/mean_terminated_length": 203.0773483276367, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02612366092701753, "epoch": 33.323232323232325, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5572985071.0, "reward": 7.7827729225158695, "reward_std": 0.28353718519210813, "rewards/combined_reward/mean": 7.78277268409729, "rewards/combined_reward/std": 2.535268986225128, "step": 54450 }, { "clip_ratio/high_max": 0.002474831450187291, "clip_ratio/high_mean": 0.0014166145037355212, "clip_ratio/low_mean": 0.0008025392246054253, "clip_ratio/low_min": 0.00020657878946318912, "clip_ratio/region_mean": 0.0022191537219138505, "completions/clipped_ratio": 0.0, "completions/max_length": 603.2, "completions/max_terminated_length": 603.2, "completions/mean_length": 201.36849365234374, "completions/mean_terminated_length": 201.36849365234374, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03490720675326884, "epoch": 33.34159779614325, "frac_reward_zero_std": 0.13750000409781932, "grad_norm": 0.005096435546875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5576058486.0, "reward": 8.190385723114014, "reward_std": 0.29357125908136367, "rewards/combined_reward/mean": 8.190385389328004, "rewards/combined_reward/std": 2.0119303703308105, "step": 54480 }, { "clip_ratio/high_max": 0.0021168420159180337, "clip_ratio/high_mean": 0.001162141673315394, "clip_ratio/low_mean": 0.0007071287819902257, "clip_ratio/low_min": 0.0002066507307366313, "clip_ratio/region_mean": 0.0018692704441491514, "completions/clipped_ratio": 0.0, "completions/max_length": 620.6, "completions/max_terminated_length": 620.6, "completions/mean_length": 204.42422332763672, "completions/mean_terminated_length": 204.42422332763672, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.029781996645033358, "epoch": 33.35996326905418, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.0341796875, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5579143635.0, "reward": 8.153506183624268, "reward_std": 0.3626144349575043, "rewards/combined_reward/mean": 8.15350570678711, "rewards/combined_reward/std": 2.3348987221717836, "step": 54510 }, { "clip_ratio/high_max": 0.002227138700739791, "clip_ratio/high_mean": 0.0012348299180303002, "clip_ratio/low_mean": 0.0008489646466235475, "clip_ratio/low_min": 0.00023254154560466608, "clip_ratio/region_mean": 0.00208379456307739, "completions/clipped_ratio": 0.0, "completions/max_length": 616.2, "completions/max_terminated_length": 616.2, "completions/mean_length": 208.94427490234375, "completions/mean_terminated_length": 208.94427490234375, "completions/min_length": 49.6, "completions/min_terminated_length": 49.6, "entropy": 0.03164950930513442, "epoch": 33.378328741965106, "frac_reward_zero_std": 0.19166667461395265, "grad_norm": 0.01202392578125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5582246141.0, "reward": 7.906573057174683, "reward_std": 0.32056449353694916, "rewards/combined_reward/mean": 7.906572675704956, "rewards/combined_reward/std": 2.4348127603530885, "step": 54540 }, { "clip_ratio/high_max": 0.0020109618878147253, "clip_ratio/high_mean": 0.0012029151495274468, "clip_ratio/low_mean": 0.0007254880978507572, "clip_ratio/low_min": 0.00020650597507483327, "clip_ratio/region_mean": 0.0019284032362823684, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 199.81302490234376, "completions/mean_terminated_length": 199.81302490234376, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.027450848013783495, "epoch": 33.396694214876035, "frac_reward_zero_std": 0.1708333384245634, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 5585313583.0, "reward": 8.083819246292114, "reward_std": 0.34448335990309714, "rewards/combined_reward/mean": 8.083818769454956, "rewards/combined_reward/std": 2.156011712551117, "step": 54570 }, { "clip_ratio/high_max": 0.0021067402568102503, "clip_ratio/high_mean": 0.0012478276221372653, "clip_ratio/low_mean": 0.0008316125306009781, "clip_ratio/low_min": 0.00023990528400948582, "clip_ratio/region_mean": 0.0020794401401265836, "completions/clipped_ratio": 0.0, "completions/max_length": 590.9, "completions/max_terminated_length": 590.9, "completions/mean_length": 202.5851593017578, "completions/mean_terminated_length": 202.5851593017578, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.030166537687182426, "epoch": 33.41505968778696, "frac_reward_zero_std": 0.1541666727513075, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5588391670.0, "reward": 8.194967651367188, "reward_std": 0.32803136110305786, "rewards/combined_reward/mean": 8.19496717453003, "rewards/combined_reward/std": 2.1837095260620116, "step": 54600 }, { "clip_ratio/high_max": 0.0019613477425688567, "clip_ratio/high_mean": 0.0010680961319546138, "clip_ratio/low_mean": 0.0009108916393112547, "clip_ratio/low_min": 0.0002989701062081925, "clip_ratio/region_mean": 0.0019789877764803047, "completions/clipped_ratio": 0.0, "completions/max_length": 558.5, "completions/max_terminated_length": 558.5, "completions/mean_length": 194.52292022705078, "completions/mean_terminated_length": 194.52292022705078, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.032398024046172695, "epoch": 33.43342516069789, "frac_reward_zero_std": 0.21666667200624942, "grad_norm": 0.015380859375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5591438798.0, "reward": 8.014452266693116, "reward_std": 0.32623865753412246, "rewards/combined_reward/mean": 8.014451932907104, "rewards/combined_reward/std": 2.295271670818329, "step": 54630 }, { "clip_ratio/high_max": 0.001460735973281165, "clip_ratio/high_mean": 0.0008514678789651953, "clip_ratio/low_mean": 0.0007712142020560956, "clip_ratio/low_min": 0.0002244710449303966, "clip_ratio/region_mean": 0.0016226820783534397, "completions/clipped_ratio": 0.0, "completions/max_length": 529.6, "completions/max_terminated_length": 529.6, "completions/mean_length": 200.20885772705077, "completions/mean_terminated_length": 200.20885772705077, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.02730349946456651, "epoch": 33.451790633608816, "frac_reward_zero_std": 0.14583333842456342, "grad_norm": 0.02197265625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5594507760.0, "reward": 7.780155992507934, "reward_std": 0.31286838799715044, "rewards/combined_reward/mean": 7.780155849456787, "rewards/combined_reward/std": 2.361257570981979, "step": 54660 }, { "clip_ratio/high_max": 0.001908383432358581, "clip_ratio/high_mean": 0.0010360995335455907, "clip_ratio/low_mean": 0.0008250141194973064, "clip_ratio/low_min": 0.00021827018669379566, "clip_ratio/region_mean": 0.0018611136523153011, "completions/clipped_ratio": 0.0, "completions/max_length": 596.9, "completions/max_terminated_length": 596.9, "completions/mean_length": 193.06068115234376, "completions/mean_terminated_length": 193.06068115234376, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.028191405845185123, "epoch": 33.470156106519745, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.054443359375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5597549273.0, "reward": 8.015360689163208, "reward_std": 0.3296437576413155, "rewards/combined_reward/mean": 8.015360355377197, "rewards/combined_reward/std": 2.3479244470596314, "step": 54690 }, { "clip_ratio/high_max": 0.0019516537776022838, "clip_ratio/high_mean": 0.001131114577583503, "clip_ratio/low_mean": 0.0006881372601128533, "clip_ratio/low_min": 0.00019104932386350507, "clip_ratio/region_mean": 0.0018192518332701487, "completions/clipped_ratio": 0.0, "completions/max_length": 563.7, "completions/max_terminated_length": 563.7, "completions/mean_length": 204.7487014770508, "completions/mean_terminated_length": 204.7487014770508, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.03014524606987834, "epoch": 33.48852157943067, "frac_reward_zero_std": 0.13750000223517417, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 5600635668.0, "reward": 8.253063297271728, "reward_std": 0.28739711493253706, "rewards/combined_reward/mean": 8.253062772750855, "rewards/combined_reward/std": 2.1735400915145875, "step": 54720 }, { "clip_ratio/high_max": 0.00183476834014679, "clip_ratio/high_mean": 0.001079621435928857, "clip_ratio/low_mean": 0.0008550581898210415, "clip_ratio/low_min": 0.00029240004829868363, "clip_ratio/region_mean": 0.0019346796121681108, "completions/clipped_ratio": 0.0, "completions/max_length": 590.9, "completions/max_terminated_length": 590.9, "completions/mean_length": 198.1408905029297, "completions/mean_terminated_length": 198.1408905029297, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028503437635178368, "epoch": 33.5068870523416, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.005706787109375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5603696689.0, "reward": 7.8142273902893065, "reward_std": 0.3006785482168198, "rewards/combined_reward/mean": 7.814227056503296, "rewards/combined_reward/std": 2.466781258583069, "step": 54750 }, { "clip_ratio/high_max": 0.0020866616289519396, "clip_ratio/high_mean": 0.0011794567715696757, "clip_ratio/low_mean": 0.0006754519701644313, "clip_ratio/low_min": 0.00016525454363242414, "clip_ratio/region_mean": 0.0018549087437956283, "completions/clipped_ratio": 0.0, "completions/max_length": 603.5, "completions/max_terminated_length": 603.5, "completions/mean_length": 188.90287017822266, "completions/mean_terminated_length": 188.90287017822266, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02867495046618084, "epoch": 33.525252525252526, "frac_reward_zero_std": 0.20000000894069672, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5606722236.0, "reward": 7.851581859588623, "reward_std": 0.2984939530491829, "rewards/combined_reward/mean": 7.851581764221192, "rewards/combined_reward/std": 2.5780961513519287, "step": 54780 }, { "clip_ratio/high_max": 0.002227853067355075, "clip_ratio/high_mean": 0.0012642884403855229, "clip_ratio/low_mean": 0.0007523351040921018, "clip_ratio/low_min": 0.00014651763024933946, "clip_ratio/region_mean": 0.0020166235524811784, "completions/clipped_ratio": 0.0, "completions/max_length": 558.4, "completions/max_terminated_length": 558.4, "completions/mean_length": 204.44870147705078, "completions/mean_terminated_length": 204.44870147705078, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02915854355475555, "epoch": 33.543617998163455, "frac_reward_zero_std": 0.15416666977107524, "grad_norm": 0.01177978515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5609807479.0, "reward": 8.199527072906495, "reward_std": 0.3086620345711708, "rewards/combined_reward/mean": 8.1995267868042, "rewards/combined_reward/std": 2.2402820229530334, "step": 54810 }, { "clip_ratio/high_max": 0.0018612732555387387, "clip_ratio/high_mean": 0.000998769694706425, "clip_ratio/low_mean": 0.000756393155522043, "clip_ratio/low_min": 0.00018323885839587699, "clip_ratio/region_mean": 0.001755162822761728, "completions/clipped_ratio": 0.0, "completions/max_length": 626.6, "completions/max_terminated_length": 626.6, "completions/mean_length": 200.75677490234375, "completions/mean_terminated_length": 200.75677490234375, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.02810770832778265, "epoch": 33.56198347107438, "frac_reward_zero_std": 0.19583333805203437, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5612878545.0, "reward": 8.089554786682129, "reward_std": 0.23770904764533043, "rewards/combined_reward/mean": 8.089554500579833, "rewards/combined_reward/std": 2.0341952204704286, "step": 54840 }, { "clip_ratio/high_max": 0.0020139824778501255, "clip_ratio/high_mean": 0.0011408469945308752, "clip_ratio/low_mean": 0.0007463826277671615, "clip_ratio/low_min": 0.00028364654329683013, "clip_ratio/region_mean": 0.0018872296432770478, "completions/clipped_ratio": 0.0, "completions/max_length": 612.1, "completions/max_terminated_length": 612.1, "completions/mean_length": 213.65443267822266, "completions/mean_terminated_length": 213.65443267822266, "completions/min_length": 50.9, "completions/min_terminated_length": 50.9, "entropy": 0.030071501868466535, "epoch": 33.58034894398531, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.0140380859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5615999138.0, "reward": 7.861095333099366, "reward_std": 0.33593177050352097, "rewards/combined_reward/mean": 7.8610950946807865, "rewards/combined_reward/std": 2.3680795550346376, "step": 54870 }, { "clip_ratio/high_max": 0.0018316848368461554, "clip_ratio/high_mean": 0.0010583619880587018, "clip_ratio/low_mean": 0.000618386810310767, "clip_ratio/low_min": 0.00017159073371052121, "clip_ratio/region_mean": 0.0016767488021287135, "completions/clipped_ratio": 0.0, "completions/max_length": 543.1, "completions/max_terminated_length": 543.1, "completions/mean_length": 212.58542175292968, "completions/mean_terminated_length": 212.58542175292968, "completions/min_length": 45.9, "completions/min_terminated_length": 45.9, "entropy": 0.02499181409366429, "epoch": 33.598714416896236, "frac_reward_zero_std": 0.15416666939854623, "grad_norm": 0.004730224609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5619115626.0, "reward": 8.067545938491822, "reward_std": 0.3691131979227066, "rewards/combined_reward/mean": 8.067545747756958, "rewards/combined_reward/std": 2.1335800766944883, "step": 54900 }, { "clip_ratio/high_max": 0.0019380639947485178, "clip_ratio/high_mean": 0.0010689686615175257, "clip_ratio/low_mean": 0.0008331897863778674, "clip_ratio/low_min": 0.00020436308550415562, "clip_ratio/region_mean": 0.0019021584400131057, "completions/clipped_ratio": 0.0, "completions/max_length": 532.6, "completions/max_terminated_length": 532.6, "completions/mean_length": 201.1596420288086, "completions/mean_terminated_length": 201.1596420288086, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.028998270879189172, "epoch": 33.617079889807165, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.005218505859375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5622188239.0, "reward": 8.017138242721558, "reward_std": 0.34965122640132906, "rewards/combined_reward/mean": 8.0171377658844, "rewards/combined_reward/std": 2.212871050834656, "step": 54930 }, { "clip_ratio/high_max": 0.002214351097063627, "clip_ratio/high_mean": 0.0012324903327680659, "clip_ratio/low_mean": 0.0007115105516277253, "clip_ratio/low_min": 0.00020433815006981604, "clip_ratio/region_mean": 0.0019440008797876847, "completions/clipped_ratio": 0.0, "completions/max_length": 597.8, "completions/max_terminated_length": 597.8, "completions/mean_length": 209.06094207763672, "completions/mean_terminated_length": 209.06094207763672, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.030014319717884062, "epoch": 33.63544536271809, "frac_reward_zero_std": 0.12500000223517418, "grad_norm": 0.01904296875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5625291193.0, "reward": 8.156234550476075, "reward_std": 0.3016998007893562, "rewards/combined_reward/mean": 8.156234216690063, "rewards/combined_reward/std": 2.2050090789794923, "step": 54960 }, { "clip_ratio/high_max": 0.001793239028484095, "clip_ratio/high_mean": 0.001016268334205961, "clip_ratio/low_mean": 0.0007428429067658726, "clip_ratio/low_min": 0.00023629889959314216, "clip_ratio/region_mean": 0.0017591112492179187, "completions/clipped_ratio": 0.0, "completions/max_length": 577.6, "completions/max_terminated_length": 577.6, "completions/mean_length": 209.52057647705078, "completions/mean_terminated_length": 209.52057647705078, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.027724968401404718, "epoch": 33.65381083562902, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5628395912.0, "reward": 8.240276956558228, "reward_std": 0.29231982976198195, "rewards/combined_reward/mean": 8.240276670455932, "rewards/combined_reward/std": 2.0049842476844786, "step": 54990 }, { "clip_ratio/high_max": 0.0018787304468181295, "clip_ratio/high_mean": 0.0010712900582196502, "clip_ratio/low_mean": 0.0008504947847541189, "clip_ratio/low_min": 0.0002996843291233139, "clip_ratio/region_mean": 0.0019217848540089713, "completions/clipped_ratio": 0.0, "completions/max_length": 545.3, "completions/max_terminated_length": 545.3, "completions/mean_length": 201.4635452270508, "completions/mean_terminated_length": 201.4635452270508, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.029508203780278564, "epoch": 33.672176308539946, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.0103759765625, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5631469692.0, "reward": 7.695294380187988, "reward_std": 0.3613412156701088, "rewards/combined_reward/mean": 7.6952942371368405, "rewards/combined_reward/std": 2.402385401725769, "step": 55020 }, { "clip_ratio/high_max": 0.0019106829863934157, "clip_ratio/high_mean": 0.0010655177393346094, "clip_ratio/low_mean": 0.0006541576323797927, "clip_ratio/low_min": 0.00014123869429264838, "clip_ratio/region_mean": 0.0017196753911169556, "completions/clipped_ratio": 0.0, "completions/max_length": 517.5, "completions/max_terminated_length": 517.5, "completions/mean_length": 196.70443115234374, "completions/mean_terminated_length": 196.70443115234374, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "entropy": 0.025456851084406176, "epoch": 33.690541781450875, "frac_reward_zero_std": 0.2000000063329935, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5634525197.0, "reward": 8.29732804298401, "reward_std": 0.2582802981138229, "rewards/combined_reward/mean": 8.29732780456543, "rewards/combined_reward/std": 2.002518367767334, "step": 55050 }, { "clip_ratio/high_max": 0.002335976681691439, "clip_ratio/high_mean": 0.0012603320472408086, "clip_ratio/low_mean": 0.0007385770039642618, "clip_ratio/low_min": 0.00020288203110491548, "clip_ratio/region_mean": 0.0019989090268306125, "completions/clipped_ratio": 0.0, "completions/max_length": 540.9, "completions/max_terminated_length": 540.9, "completions/mean_length": 209.56094207763672, "completions/mean_terminated_length": 209.56094207763672, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.029919865153109033, "epoch": 33.7089072543618, "frac_reward_zero_std": 0.12500000447034837, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5637630071.0, "reward": 8.16792426109314, "reward_std": 0.3246514111757278, "rewards/combined_reward/mean": 8.167923974990845, "rewards/combined_reward/std": 2.0722890734672545, "step": 55080 }, { "clip_ratio/high_max": 0.0020492269371364578, "clip_ratio/high_mean": 0.0012185355257922008, "clip_ratio/low_mean": 0.0008151164811958249, "clip_ratio/low_min": 0.0002676829278546696, "clip_ratio/region_mean": 0.0020336520101409405, "completions/clipped_ratio": 0.0, "completions/max_length": 562.2, "completions/max_terminated_length": 562.2, "completions/mean_length": 212.37266387939454, "completions/mean_terminated_length": 212.37266387939454, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.030300926494722564, "epoch": 33.72727272727273, "frac_reward_zero_std": 0.14583333916962146, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5640745742.0, "reward": 8.124664926528931, "reward_std": 0.3318124294281006, "rewards/combined_reward/mean": 8.124664688110352, "rewards/combined_reward/std": 2.2505564093589783, "step": 55110 }, { "clip_ratio/high_max": 0.001963857840746641, "clip_ratio/high_mean": 0.0011122523671171318, "clip_ratio/low_mean": 0.0008026933275687043, "clip_ratio/low_min": 0.0002327394899718153, "clip_ratio/region_mean": 0.0019149457094802833, "completions/clipped_ratio": 0.0, "completions/max_length": 582.3, "completions/max_terminated_length": 582.3, "completions/mean_length": 211.13490142822266, "completions/mean_terminated_length": 211.13490142822266, "completions/min_length": 48.2, "completions/min_terminated_length": 48.2, "entropy": 0.027502090359727543, "epoch": 33.745638200183656, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5643856660.0, "reward": 8.051837062835693, "reward_std": 0.3531241253018379, "rewards/combined_reward/mean": 8.051836824417114, "rewards/combined_reward/std": 2.3728591442108153, "step": 55140 }, { "clip_ratio/high_max": 0.0020180095477068485, "clip_ratio/high_mean": 0.0011201041132153477, "clip_ratio/low_mean": 0.0007096820388445243, "clip_ratio/low_min": 0.00014913759514456615, "clip_ratio/region_mean": 0.0018297861558191169, "completions/clipped_ratio": 0.0, "completions/max_length": 564.3, "completions/max_terminated_length": 564.3, "completions/mean_length": 206.09349365234374, "completions/mean_terminated_length": 206.09349365234374, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.027931311850746474, "epoch": 33.764003673094585, "frac_reward_zero_std": 0.12500000335276126, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5646948219.0, "reward": 8.189739990234376, "reward_std": 0.2928040012717247, "rewards/combined_reward/mean": 8.189739656448364, "rewards/combined_reward/std": 2.1168742895126345, "step": 55170 }, { "clip_ratio/high_max": 0.0016992741099481161, "clip_ratio/high_mean": 0.0010157646116567775, "clip_ratio/low_mean": 0.0007838694849245561, "clip_ratio/low_min": 0.0002190236969909165, "clip_ratio/region_mean": 0.0017996341096780573, "completions/clipped_ratio": 0.0, "completions/max_length": 576.2, "completions/max_terminated_length": 576.2, "completions/mean_length": 209.0682342529297, "completions/mean_terminated_length": 209.0682342529297, "completions/min_length": 48.8, "completions/min_terminated_length": 48.8, "entropy": 0.02721629394994428, "epoch": 33.78236914600551, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.006134033203125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5650051201.0, "reward": 8.183662700653077, "reward_std": 0.27606791257858276, "rewards/combined_reward/mean": 8.183662271499633, "rewards/combined_reward/std": 2.1121095299720762, "step": 55200 }, { "clip_ratio/high_max": 0.002406357085177054, "clip_ratio/high_mean": 0.001333151305273835, "clip_ratio/low_mean": 0.0007230026464336939, "clip_ratio/low_min": 0.0001679556466115173, "clip_ratio/region_mean": 0.002056153966016912, "completions/clipped_ratio": 0.0, "completions/max_length": 543.6, "completions/max_terminated_length": 543.6, "completions/mean_length": 201.69609832763672, "completions/mean_terminated_length": 201.69609832763672, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.030037839245051146, "epoch": 33.80073461891644, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.0203857421875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5653125874.0, "reward": 8.08250937461853, "reward_std": 0.29264564588665964, "rewards/combined_reward/mean": 8.082508945465088, "rewards/combined_reward/std": 2.2413264989852903, "step": 55230 }, { "clip_ratio/high_max": 0.0019729017571080477, "clip_ratio/high_mean": 0.0010344373261129173, "clip_ratio/low_mean": 0.0008152913425874431, "clip_ratio/low_min": 0.00023278481943028357, "clip_ratio/region_mean": 0.00184972866360719, "completions/clipped_ratio": 0.0, "completions/max_length": 550.7, "completions/max_terminated_length": 550.7, "completions/mean_length": 198.80130767822266, "completions/mean_terminated_length": 198.80130767822266, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.029626329119006794, "epoch": 33.819100091827366, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5656189431.0, "reward": 7.923173952102661, "reward_std": 0.2557866141200066, "rewards/combined_reward/mean": 7.923173666000366, "rewards/combined_reward/std": 2.438519871234894, "step": 55260 }, { "clip_ratio/high_max": 0.0028068634230294263, "clip_ratio/high_mean": 0.001436022183164217, "clip_ratio/low_mean": 0.0008087121065424677, "clip_ratio/low_min": 0.00022029712296595488, "clip_ratio/region_mean": 0.0022447342974677062, "completions/clipped_ratio": 0.0, "completions/max_length": 574.5, "completions/max_terminated_length": 574.5, "completions/mean_length": 204.88359832763672, "completions/mean_terminated_length": 204.88359832763672, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.0325889655854553, "epoch": 33.837465564738295, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.006561279296875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5659276344.0, "reward": 8.226804494857788, "reward_std": 0.3131843373179436, "rewards/combined_reward/mean": 8.226804447174072, "rewards/combined_reward/std": 2.0587970852851867, "step": 55290 }, { "clip_ratio/high_max": 0.0021953954856144264, "clip_ratio/high_mean": 0.001228934970034364, "clip_ratio/low_mean": 0.0007159307666976625, "clip_ratio/low_min": 0.0001993642348679714, "clip_ratio/region_mean": 0.001944865735519367, "completions/clipped_ratio": 0.0, "completions/max_length": 577.2, "completions/max_terminated_length": 577.2, "completions/mean_length": 204.72474212646483, "completions/mean_terminated_length": 204.72474212646483, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.028136536339297892, "epoch": 33.85583103764922, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.020263671875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5662362647.0, "reward": 8.260259342193603, "reward_std": 0.2787884630262852, "rewards/combined_reward/mean": 8.260258913040161, "rewards/combined_reward/std": 2.0373384833335875, "step": 55320 }, { "clip_ratio/high_max": 0.0019740135066967923, "clip_ratio/high_mean": 0.0011529699428744304, "clip_ratio/low_mean": 0.0009431373814247006, "clip_ratio/low_min": 0.00034278831420427497, "clip_ratio/region_mean": 0.0020961073275733116, "completions/clipped_ratio": 0.0, "completions/max_length": 590.8, "completions/max_terminated_length": 590.8, "completions/mean_length": 209.7362045288086, "completions/mean_terminated_length": 209.7362045288086, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "entropy": 0.03156464053317905, "epoch": 33.87419651056015, "frac_reward_zero_std": 0.15416667237877846, "grad_norm": 0.025390625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5665468194.0, "reward": 8.310786581039428, "reward_std": 0.3350495234131813, "rewards/combined_reward/mean": 8.310786247253418, "rewards/combined_reward/std": 1.9580763936042787, "step": 55350 }, { "clip_ratio/high_max": 0.0016031468815829915, "clip_ratio/high_mean": 0.0009806171504654533, "clip_ratio/low_mean": 0.000747662424206889, "clip_ratio/low_min": 0.00018816636268941995, "clip_ratio/region_mean": 0.001728279545204714, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 207.79635925292968, "completions/mean_terminated_length": 207.79635925292968, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.0266384352153788, "epoch": 33.892561983471076, "frac_reward_zero_std": 0.17083333656191826, "grad_norm": 0.010009765625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5668566292.0, "reward": 8.133120441436768, "reward_std": 0.31846220046281815, "rewards/combined_reward/mean": 8.133120346069337, "rewards/combined_reward/std": 2.1208059668540953, "step": 55380 }, { "clip_ratio/high_max": 0.0015174941096726495, "clip_ratio/high_mean": 0.000839640426056576, "clip_ratio/low_mean": 0.0008121364001757078, "clip_ratio/low_min": 0.0002182150000104836, "clip_ratio/region_mean": 0.0016517768269598795, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 201.34193267822266, "completions/mean_terminated_length": 201.34193267822266, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.027610086331454415, "epoch": 33.910927456382005, "frac_reward_zero_std": 0.17916666977107526, "grad_norm": 0.00732421875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5671639605.0, "reward": 8.071895837783813, "reward_std": 0.34017928689718246, "rewards/combined_reward/mean": 8.071895122528076, "rewards/combined_reward/std": 2.033528757095337, "step": 55410 }, { "clip_ratio/high_max": 0.0019774294928841603, "clip_ratio/high_mean": 0.0011582538791117258, "clip_ratio/low_mean": 0.0007412216312028856, "clip_ratio/low_min": 0.00021902242660871708, "clip_ratio/region_mean": 0.0018994755213498139, "completions/clipped_ratio": 0.0, "completions/max_length": 686.1, "completions/max_terminated_length": 686.1, "completions/mean_length": 202.7010467529297, "completions/mean_terminated_length": 202.7010467529297, "completions/min_length": 44.5, "completions/min_terminated_length": 44.5, "entropy": 0.03085805553322037, "epoch": 33.92929292929293, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.007568359375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5674718137.0, "reward": 8.191962480545044, "reward_std": 0.2992715373635292, "rewards/combined_reward/mean": 8.191962099075317, "rewards/combined_reward/std": 2.0791945815086366, "step": 55440 }, { "clip_ratio/high_max": 0.0017501448858335304, "clip_ratio/high_mean": 0.0009741777947056107, "clip_ratio/low_mean": 0.0008276064868065684, "clip_ratio/low_min": 0.0002474926429082795, "clip_ratio/region_mean": 0.001801784275448881, "completions/clipped_ratio": 0.0, "completions/max_length": 550.3, "completions/max_terminated_length": 550.3, "completions/mean_length": 201.29896240234376, "completions/mean_terminated_length": 201.29896240234376, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.026075116265565156, "epoch": 33.94765840220386, "frac_reward_zero_std": 0.22916667014360428, "grad_norm": 0.006591796875, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 5677791285.0, "reward": 7.968142366409301, "reward_std": 0.26415332183241846, "rewards/combined_reward/mean": 7.968142366409301, "rewards/combined_reward/std": 2.401633071899414, "step": 55470 }, { "clip_ratio/high_max": 0.0020423895342294904, "clip_ratio/high_mean": 0.0011581721346980581, "clip_ratio/low_mean": 0.0007476277368065591, "clip_ratio/low_min": 0.0002466073579853401, "clip_ratio/region_mean": 0.0019057998763552556, "completions/clipped_ratio": 0.0, "completions/max_length": 647.3, "completions/max_terminated_length": 647.3, "completions/mean_length": 200.61510772705077, "completions/mean_terminated_length": 200.61510772705077, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02640295820310712, "epoch": 33.966023875114786, "frac_reward_zero_std": 0.2041666701436043, "grad_norm": 0.00433349609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5680861807.0, "reward": 7.887787818908691, "reward_std": 0.25855070203542707, "rewards/combined_reward/mean": 7.887787485122681, "rewards/combined_reward/std": 2.5121782422065735, "step": 55500 }, { "clip_ratio/high_max": 0.0017011852881599527, "clip_ratio/high_mean": 0.00099207558935935, "clip_ratio/low_mean": 0.0006552094333528658, "clip_ratio/low_min": 0.00018479479210024388, "clip_ratio/region_mean": 0.001647285003370295, "completions/clipped_ratio": 0.0, "completions/max_length": 671.4, "completions/max_terminated_length": 671.4, "completions/mean_length": 203.73047332763673, "completions/mean_terminated_length": 203.73047332763673, "completions/min_length": 38.9, "completions/min_terminated_length": 38.9, "entropy": 0.024871965885783234, "epoch": 33.984389348025715, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.0126953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5683944292.0, "reward": 8.142454624176025, "reward_std": 0.27818151712417605, "rewards/combined_reward/mean": 8.142454576492309, "rewards/combined_reward/std": 2.042375034093857, "step": 55530 }, { "clip_ratio/high_max": 0.001728608931384248, "clip_ratio/high_mean": 0.00102307827516791, "clip_ratio/low_mean": 0.0007509304843671223, "clip_ratio/low_min": 0.00022897580910481697, "clip_ratio/region_mean": 0.0017740087699122176, "completions/clipped_ratio": 0.0, "completions/max_length": 605.1818181818181, "completions/max_terminated_length": 605.1818181818181, "completions/mean_length": 208.6351845481179, "completions/mean_terminated_length": 208.6351845481179, "completions/min_length": 49.36363636363637, "completions/min_terminated_length": 49.36363636363637, "entropy": 0.02670742899606951, "epoch": 34.003060912151824, "frac_reward_zero_std": 0.1515151560306549, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5687355743.0, "reward": 8.2819993279197, "reward_std": 0.30404620278965344, "rewards/combined_reward/mean": 8.28199915452437, "rewards/combined_reward/std": 2.2156017260117964, "step": 55560 }, { "clip_ratio/high_max": 0.0026732069917973908, "clip_ratio/high_mean": 0.0014713655200466746, "clip_ratio/low_mean": 0.0010596706942427167, "clip_ratio/low_min": 0.00024106059160355168, "clip_ratio/region_mean": 0.0025310362136224283, "completions/clipped_ratio": 0.0, "completions/max_length": 484.3, "completions/max_terminated_length": 484.3, "completions/mean_length": 192.8026092529297, "completions/mean_terminated_length": 192.8026092529297, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.031521054233113924, "epoch": 34.021426385062746, "frac_reward_zero_std": 0.17083333507180215, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 5690396265.0, "reward": 7.993099927902222, "reward_std": 0.27941213846206664, "rewards/combined_reward/mean": 7.993099784851074, "rewards/combined_reward/std": 2.272440505027771, "step": 55590 }, { "clip_ratio/high_max": 0.0025472453424299603, "clip_ratio/high_mean": 0.0015591196845586334, "clip_ratio/low_mean": 0.0009961415328386162, "clip_ratio/low_min": 0.0002718505184020614, "clip_ratio/region_mean": 0.0025552612380124627, "completions/clipped_ratio": 0.0, "completions/max_length": 523.1, "completions/max_terminated_length": 523.1, "completions/mean_length": 213.97083892822266, "completions/mean_terminated_length": 213.97083892822266, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.031207815798309942, "epoch": 34.039791857973675, "frac_reward_zero_std": 0.15416667237877846, "grad_norm": 0.01275634765625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5693518073.0, "reward": 8.10495958328247, "reward_std": 0.32276115715503695, "rewards/combined_reward/mean": 8.104959440231323, "rewards/combined_reward/std": 2.209471952915192, "step": 55620 }, { "clip_ratio/high_max": 0.0023542753876730177, "clip_ratio/high_mean": 0.0013955798618553671, "clip_ratio/low_mean": 0.0008434797192118519, "clip_ratio/low_min": 0.00017541121414978988, "clip_ratio/region_mean": 0.0022390595976806557, "completions/clipped_ratio": 0.0, "completions/max_length": 533.1, "completions/max_terminated_length": 533.1, "completions/mean_length": 200.35547485351563, "completions/mean_terminated_length": 200.35547485351563, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.027773556097721062, "epoch": 34.058157330884605, "frac_reward_zero_std": 0.13333333767950534, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5696587598.0, "reward": 8.161719942092896, "reward_std": 0.2832194253802299, "rewards/combined_reward/mean": 8.161719703674317, "rewards/combined_reward/std": 2.0961719155311584, "step": 55650 }, { "clip_ratio/high_max": 0.0021433344333975884, "clip_ratio/high_mean": 0.001289520617259162, "clip_ratio/low_mean": 0.0008692714024315743, "clip_ratio/low_min": 0.0002620418437193924, "clip_ratio/region_mean": 0.0021587919817344908, "completions/clipped_ratio": 0.0, "completions/max_length": 560.9, "completions/max_terminated_length": 560.9, "completions/mean_length": 196.91484680175782, "completions/mean_terminated_length": 196.91484680175782, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.027963453733051816, "epoch": 34.076522803795534, "frac_reward_zero_std": 0.0958333358168602, "grad_norm": 0.0147705078125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5699643911.0, "reward": 8.262535858154298, "reward_std": 0.3338841527700424, "rewards/combined_reward/mean": 8.262535810470581, "rewards/combined_reward/std": 2.0935351610183717, "step": 55680 }, { "clip_ratio/high_max": 0.0023053893564792816, "clip_ratio/high_mean": 0.0013544310785315852, "clip_ratio/low_mean": 0.0008921153520835408, "clip_ratio/low_min": 0.00023218380277588343, "clip_ratio/region_mean": 0.0022465464178822005, "completions/clipped_ratio": 0.0, "completions/max_length": 584.6, "completions/max_terminated_length": 584.6, "completions/mean_length": 207.8260498046875, "completions/mean_terminated_length": 207.8260498046875, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.028549739082033437, "epoch": 34.094888276706456, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0233154296875, "learning_rate": 5e-06, "loss": 0.0004, "num_tokens": 5702742123.0, "reward": 8.083037376403809, "reward_std": 0.28980927243828775, "rewards/combined_reward/mean": 8.083036994934082, "rewards/combined_reward/std": 2.074861204624176, "step": 55710 }, { "clip_ratio/high_max": 0.002485750627965899, "clip_ratio/high_mean": 0.0013577688869190752, "clip_ratio/low_mean": 0.0009350848108927797, "clip_ratio/low_min": 0.00029263712543373306, "clip_ratio/region_mean": 0.0022928537160623818, "completions/clipped_ratio": 0.0, "completions/max_length": 773.7, "completions/max_terminated_length": 773.7, "completions/mean_length": 214.61068267822264, "completions/mean_terminated_length": 214.61068267822264, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.030182609458764394, "epoch": 34.113253749617385, "frac_reward_zero_std": 0.15416666865348816, "grad_norm": 0.01611328125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 5705866388.0, "reward": 7.876918458938599, "reward_std": 0.31593531370162964, "rewards/combined_reward/mean": 7.8769182682037355, "rewards/combined_reward/std": 2.456935095787048, "step": 55740 }, { "clip_ratio/high_max": 0.0024161315552191807, "clip_ratio/high_mean": 0.001427672800855362, "clip_ratio/low_mean": 0.0009836546014412306, "clip_ratio/low_min": 0.00025457237449397023, "clip_ratio/region_mean": 0.0024113274055707736, "completions/clipped_ratio": 0.0, "completions/max_length": 672.5, "completions/max_terminated_length": 672.5, "completions/mean_length": 196.05521392822266, "completions/mean_terminated_length": 196.05521392822266, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03088177617173642, "epoch": 34.131619222528315, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.00897216796875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5708919400.0, "reward": 8.054835605621339, "reward_std": 0.2880806714296341, "rewards/combined_reward/mean": 8.05483546257019, "rewards/combined_reward/std": 2.3168464303016663, "step": 55770 }, { "clip_ratio/high_max": 0.0027034626837121323, "clip_ratio/high_mean": 0.0014123581363188957, "clip_ratio/low_mean": 0.0008804482845638025, "clip_ratio/low_min": 0.0001420099977015828, "clip_ratio/region_mean": 0.0022928064441657624, "completions/clipped_ratio": 0.0, "completions/max_length": 568.9, "completions/max_terminated_length": 568.9, "completions/mean_length": 198.64896545410156, "completions/mean_terminated_length": 198.64896545410156, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.02820287096935014, "epoch": 34.149984695439244, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 5711982372.0, "reward": 8.003286600112915, "reward_std": 0.3612269774079323, "rewards/combined_reward/mean": 8.003286457061767, "rewards/combined_reward/std": 2.251749002933502, "step": 55800 }, { "clip_ratio/high_max": 0.002368846881048133, "clip_ratio/high_mean": 0.0013293766575467696, "clip_ratio/low_mean": 0.000952338947536191, "clip_ratio/low_min": 0.00023654832742370975, "clip_ratio/region_mean": 0.002281715621938929, "completions/clipped_ratio": 0.0, "completions/max_length": 560.1, "completions/max_terminated_length": 560.1, "completions/mean_length": 209.32891387939452, "completions/mean_terminated_length": 209.32891387939452, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.0274117482205232, "epoch": 34.168350168350166, "frac_reward_zero_std": 0.14583333767950535, "grad_norm": 0.011474609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5715086355.0, "reward": 8.033217239379884, "reward_std": 0.31995334178209306, "rewards/combined_reward/mean": 8.033217191696167, "rewards/combined_reward/std": 2.3355359077453612, "step": 55830 }, { "clip_ratio/high_max": 0.002051638353441376, "clip_ratio/high_mean": 0.0012122653252542174, "clip_ratio/low_mean": 0.0007501259009586647, "clip_ratio/low_min": 0.00018542154696964038, "clip_ratio/region_mean": 0.0019623912309422544, "completions/clipped_ratio": 0.0, "completions/max_length": 551.9, "completions/max_terminated_length": 551.9, "completions/mean_length": 201.8726608276367, "completions/mean_terminated_length": 201.8726608276367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.025463503474990528, "epoch": 34.186715641261095, "frac_reward_zero_std": 0.17916666828095912, "grad_norm": 0.00689697265625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5718161706.0, "reward": 8.13477864265442, "reward_std": 0.23972296640276908, "rewards/combined_reward/mean": 8.134778594970703, "rewards/combined_reward/std": 2.235758030414581, "step": 55860 }, { "clip_ratio/high_max": 0.002650398885695419, "clip_ratio/high_mean": 0.001499573119872366, "clip_ratio/low_mean": 0.0009010084808930212, "clip_ratio/low_min": 0.00025985124011640437, "clip_ratio/region_mean": 0.002400581586698536, "completions/clipped_ratio": 0.0, "completions/max_length": 561.6, "completions/max_terminated_length": 561.6, "completions/mean_length": 193.88412017822264, "completions/mean_terminated_length": 193.88412017822264, "completions/min_length": 39.5, "completions/min_terminated_length": 39.5, "entropy": 0.028524729143828153, "epoch": 34.205081114172025, "frac_reward_zero_std": 0.2125000089406967, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5721206381.0, "reward": 8.07986044883728, "reward_std": 0.24194100350141526, "rewards/combined_reward/mean": 8.079860258102418, "rewards/combined_reward/std": 2.1853311419487, "step": 55890 }, { "clip_ratio/high_max": 0.0023098354756560485, "clip_ratio/high_mean": 0.0013963596141062832, "clip_ratio/low_mean": 0.0010025793770182645, "clip_ratio/low_min": 0.00028634992074027345, "clip_ratio/region_mean": 0.0023989389854250477, "completions/clipped_ratio": 0.0, "completions/max_length": 696.1, "completions/max_terminated_length": 696.1, "completions/mean_length": 196.95052642822264, "completions/mean_terminated_length": 196.95052642822264, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.030541519114437202, "epoch": 34.223446587082954, "frac_reward_zero_std": 0.17916667126119137, "grad_norm": 0.0146484375, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5724262831.0, "reward": 8.034900856018066, "reward_std": 0.2744539603590965, "rewards/combined_reward/mean": 8.034900522232055, "rewards/combined_reward/std": 2.455891501903534, "step": 55920 }, { "clip_ratio/high_max": 0.002358538083353778, "clip_ratio/high_mean": 0.0013668772577754378, "clip_ratio/low_mean": 0.0010001290330061844, "clip_ratio/low_min": 0.0002609395046116939, "clip_ratio/region_mean": 0.002367006278655026, "completions/clipped_ratio": 0.0, "completions/max_length": 580.9, "completions/max_terminated_length": 580.9, "completions/mean_length": 197.8195358276367, "completions/mean_terminated_length": 197.8195358276367, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02898557803904017, "epoch": 34.241812059993876, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.01373291015625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5727322618.0, "reward": 8.161690616607666, "reward_std": 0.28278072029352186, "rewards/combined_reward/mean": 8.16169033050537, "rewards/combined_reward/std": 2.0141226172447206, "step": 55950 }, { "clip_ratio/high_max": 0.002311436053181145, "clip_ratio/high_mean": 0.001391662172075788, "clip_ratio/low_mean": 0.0007859476738910113, "clip_ratio/low_min": 0.0002091340174956713, "clip_ratio/region_mean": 0.0021776098277769052, "completions/clipped_ratio": 0.0, "completions/max_length": 609.3, "completions/max_terminated_length": 609.3, "completions/mean_length": 202.8963607788086, "completions/mean_terminated_length": 202.8963607788086, "completions/min_length": 42.8, "completions/min_terminated_length": 42.8, "entropy": 0.027354116380835573, "epoch": 34.260177532904805, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.006439208984375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5730401900.0, "reward": 8.093315601348877, "reward_std": 0.2869091659784317, "rewards/combined_reward/mean": 8.093315601348877, "rewards/combined_reward/std": 2.2146299719810485, "step": 55980 }, { "clip_ratio/high_max": 0.0019267230527475476, "clip_ratio/high_mean": 0.001242032929440029, "clip_ratio/low_mean": 0.0005854107886989369, "clip_ratio/low_min": 0.00010582264076219872, "clip_ratio/region_mean": 0.0018274436995852738, "completions/clipped_ratio": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 195.70703506469727, "completions/mean_terminated_length": 195.70703506469727, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029534394200891257, "epoch": 34.27793082338537, "frac_reward_zero_std": 0.2187500074505806, "grad_norm": 0.0118408203125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 5733450621.0, "reward": 8.019140601158142, "reward_std": 0.25527874380350113, "rewards/combined_reward/mean": 8.019140243530273, "rewards/combined_reward/std": 2.3821109235286713, "step": 56010 }, { "clip_ratio/high_max": 0.0019546210756137346, "clip_ratio/high_mean": 0.0011076878795089821, "clip_ratio/low_mean": 0.0008218428737260789, "clip_ratio/low_min": 0.0002994451987130257, "clip_ratio/region_mean": 0.0019295307642702634, "completions/clipped_ratio": 0.0, "completions/max_length": 609.8, "completions/max_terminated_length": 609.8, "completions/mean_length": 210.6471405029297, "completions/mean_terminated_length": 210.6471405029297, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.032214853027835486, "epoch": 34.2962962962963, "frac_reward_zero_std": 0.1625000059604645, "grad_norm": 0.011962890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5736559666.0, "reward": 8.048356628417968, "reward_std": 0.3301005601882935, "rewards/combined_reward/mean": 8.048356342315675, "rewards/combined_reward/std": 2.142339599132538, "step": 56040 }, { "clip_ratio/high_max": 0.0015347513961993778, "clip_ratio/high_mean": 0.0008913945863848008, "clip_ratio/low_mean": 0.0006829647403113389, "clip_ratio/low_min": 0.0002182724999632531, "clip_ratio/region_mean": 0.0015743593399141294, "completions/clipped_ratio": 0.0, "completions/max_length": 622.2, "completions/max_terminated_length": 622.2, "completions/mean_length": 203.55521392822266, "completions/mean_terminated_length": 203.55521392822266, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.025038252662246425, "epoch": 34.31466176920723, "frac_reward_zero_std": 0.17083333879709245, "grad_norm": 0.006866455078125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5739641478.0, "reward": 7.887470722198486, "reward_std": 0.2805391579866409, "rewards/combined_reward/mean": 7.887470674514771, "rewards/combined_reward/std": 2.263685965538025, "step": 56070 }, { "clip_ratio/high_max": 0.002085936601118495, "clip_ratio/high_mean": 0.0012481668081212169, "clip_ratio/low_mean": 0.0008187696811849794, "clip_ratio/low_min": 0.00017584123537138415, "clip_ratio/region_mean": 0.00206693647778593, "completions/clipped_ratio": 0.0, "completions/max_length": 608.6, "completions/max_terminated_length": 608.6, "completions/mean_length": 204.23828735351563, "completions/mean_terminated_length": 204.23828735351563, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03010599670621256, "epoch": 34.33302724211815, "frac_reward_zero_std": 0.15833333991467952, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5742725913.0, "reward": 8.191229867935181, "reward_std": 0.3242354914546013, "rewards/combined_reward/mean": 8.191229677200317, "rewards/combined_reward/std": 2.1469292759895326, "step": 56100 }, { "clip_ratio/high_max": 0.0021225731892627665, "clip_ratio/high_mean": 0.001237418034482592, "clip_ratio/low_mean": 0.0008563169946379882, "clip_ratio/low_min": 0.000264872568610978, "clip_ratio/region_mean": 0.00209373502342108, "completions/clipped_ratio": 0.0, "completions/max_length": 566.4, "completions/max_terminated_length": 566.4, "completions/mean_length": 206.6955795288086, "completions/mean_terminated_length": 206.6955795288086, "completions/min_length": 47.7, "completions/min_terminated_length": 47.7, "entropy": 0.03239409706244866, "epoch": 34.35139271502908, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.016357421875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 5745819784.0, "reward": 8.074338293075561, "reward_std": 0.29734392315149305, "rewards/combined_reward/mean": 8.074338245391846, "rewards/combined_reward/std": 2.238077175617218, "step": 56130 }, { "clip_ratio/high_max": 0.002360181946035785, "clip_ratio/high_mean": 0.00128780593877309, "clip_ratio/low_mean": 0.0008868601585466725, "clip_ratio/low_min": 0.0002454669178405311, "clip_ratio/region_mean": 0.0021746661150245927, "completions/clipped_ratio": 0.0, "completions/max_length": 575.3, "completions/max_terminated_length": 575.3, "completions/mean_length": 195.7463607788086, "completions/mean_terminated_length": 195.7463607788086, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.03234430553081135, "epoch": 34.36975818794001, "frac_reward_zero_std": 0.17916667237877845, "grad_norm": 0.005859375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5748871610.0, "reward": 8.058726978302001, "reward_std": 0.32717857509851456, "rewards/combined_reward/mean": 8.058726739883422, "rewards/combined_reward/std": 2.3746358036994932, "step": 56160 }, { "clip_ratio/high_max": 0.0018236734761255017, "clip_ratio/high_mean": 0.0010269889012609686, "clip_ratio/low_mean": 0.0006914450271627478, "clip_ratio/low_min": 0.00016870910670453061, "clip_ratio/region_mean": 0.001718433938610057, "completions/clipped_ratio": 0.0, "completions/max_length": 659.7, "completions/max_terminated_length": 659.7, "completions/mean_length": 216.46068267822267, "completions/mean_terminated_length": 216.46068267822267, "completions/min_length": 44.7, "completions/min_terminated_length": 44.7, "entropy": 0.02839865400455892, "epoch": 34.38812366085094, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5752002979.0, "reward": 7.926990127563476, "reward_std": 0.34762597680091856, "rewards/combined_reward/mean": 7.926989936828614, "rewards/combined_reward/std": 2.407870388031006, "step": 56190 }, { "clip_ratio/high_max": 0.002024821232286437, "clip_ratio/high_mean": 0.001203983951806246, "clip_ratio/low_mean": 0.0007312437862007452, "clip_ratio/low_min": 0.00019266276528166296, "clip_ratio/region_mean": 0.0019352277061746765, "completions/clipped_ratio": 0.0, "completions/max_length": 591.3, "completions/max_terminated_length": 591.3, "completions/mean_length": 197.39010925292968, "completions/mean_terminated_length": 197.39010925292968, "completions/min_length": 48.5, "completions/min_terminated_length": 48.5, "entropy": 0.029414448452492557, "epoch": 34.40648913376186, "frac_reward_zero_std": 0.17083334177732468, "grad_norm": 0.00677490234375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5755061117.0, "reward": 8.302323579788208, "reward_std": 0.25265097320079805, "rewards/combined_reward/mean": 8.302322959899902, "rewards/combined_reward/std": 1.9746845841407776, "step": 56220 }, { "clip_ratio/high_max": 0.002025177260414542, "clip_ratio/high_mean": 0.0011237827351578745, "clip_ratio/low_mean": 0.0008846155600622296, "clip_ratio/low_min": 0.00021739346072232972, "clip_ratio/region_mean": 0.002008398316684179, "completions/clipped_ratio": 0.0, "completions/max_length": 600.5, "completions/max_terminated_length": 600.5, "completions/mean_length": 198.53854522705078, "completions/mean_terminated_length": 198.53854522705078, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.03302138967749973, "epoch": 34.42485460667279, "frac_reward_zero_std": 0.2041666738688946, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5758123665.0, "reward": 7.913648843765259, "reward_std": 0.32894157618284225, "rewards/combined_reward/mean": 7.913648796081543, "rewards/combined_reward/std": 2.4904292702674864, "step": 56250 }, { "clip_ratio/high_max": 0.002292912220097302, "clip_ratio/high_mean": 0.0012209201431687688, "clip_ratio/low_mean": 0.0008025133331102552, "clip_ratio/low_min": 0.0002271066116615354, "clip_ratio/region_mean": 0.0020234334656076196, "completions/clipped_ratio": 0.0, "completions/max_length": 567.3, "completions/max_terminated_length": 567.3, "completions/mean_length": 194.39010620117188, "completions/mean_terminated_length": 194.39010620117188, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029141099642341337, "epoch": 34.44322007958372, "frac_reward_zero_std": 0.1750000048428774, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5761170283.0, "reward": 8.069217824935913, "reward_std": 0.2811395779252052, "rewards/combined_reward/mean": 8.069217538833618, "rewards/combined_reward/std": 2.204007124900818, "step": 56280 }, { "clip_ratio/high_max": 0.0018733921555394773, "clip_ratio/high_mean": 0.0010460843420635987, "clip_ratio/low_mean": 0.0007859872018646759, "clip_ratio/low_min": 0.00015126495636650362, "clip_ratio/region_mean": 0.0018320715389563702, "completions/clipped_ratio": 0.0, "completions/max_length": 544.6, "completions/max_terminated_length": 544.6, "completions/mean_length": 197.94219207763672, "completions/mean_terminated_length": 197.94219207763672, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.02958859681772689, "epoch": 34.46158555249465, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.01263427734375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5764230541.0, "reward": 7.85839147567749, "reward_std": 0.3679980367422104, "rewards/combined_reward/mean": 7.858391046524048, "rewards/combined_reward/std": 2.350894045829773, "step": 56310 }, { "clip_ratio/high_max": 0.0013370519850771719, "clip_ratio/high_mean": 0.0007525966032820482, "clip_ratio/low_mean": 0.0008601410356883813, "clip_ratio/low_min": 0.0002910082689292418, "clip_ratio/region_mean": 0.0016127376244791474, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 198.14557647705078, "completions/mean_terminated_length": 198.14557647705078, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.025539818949376545, "epoch": 34.47995102540557, "frac_reward_zero_std": 0.20416667312383652, "grad_norm": 0.007659912109375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5767291580.0, "reward": 8.211501502990723, "reward_std": 0.26516171395778654, "rewards/combined_reward/mean": 8.211501169204713, "rewards/combined_reward/std": 2.239084529876709, "step": 56340 }, { "clip_ratio/high_max": 0.0021713283504747475, "clip_ratio/high_mean": 0.0011580263131084696, "clip_ratio/low_mean": 0.0008332281535937606, "clip_ratio/low_min": 0.0002105451744379631, "clip_ratio/region_mean": 0.0019912544630642515, "completions/clipped_ratio": 0.0, "completions/max_length": 573.9, "completions/max_terminated_length": 573.9, "completions/mean_length": 201.20807800292968, "completions/mean_terminated_length": 201.20807800292968, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.03135329325062533, "epoch": 34.4983164983165, "frac_reward_zero_std": 0.1750000026077032, "grad_norm": 0.0101318359375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5770364379.0, "reward": 8.059273719787598, "reward_std": 0.2961281254887581, "rewards/combined_reward/mean": 8.059273433685302, "rewards/combined_reward/std": 2.321204197406769, "step": 56370 }, { "clip_ratio/high_max": 0.002113842195831239, "clip_ratio/high_mean": 0.0011578583867958513, "clip_ratio/low_mean": 0.0007001992064033403, "clip_ratio/low_min": 0.00020845096854221385, "clip_ratio/region_mean": 0.0018580575694310635, "completions/clipped_ratio": 0.0, "completions/max_length": 570.4, "completions/max_terminated_length": 570.4, "completions/mean_length": 191.9119857788086, "completions/mean_terminated_length": 191.9119857788086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.030505789117887615, "epoch": 34.51668197122743, "frac_reward_zero_std": 0.17500000745058059, "grad_norm": 0.0045166015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5773401481.0, "reward": 7.816453695297241, "reward_std": 0.3234464704990387, "rewards/combined_reward/mean": 7.816453456878662, "rewards/combined_reward/std": 2.563649046421051, "step": 56400 }, { "clip_ratio/high_max": 0.0017794059218431356, "clip_ratio/high_mean": 0.001008089482638752, "clip_ratio/low_mean": 0.0007769719028146938, "clip_ratio/low_min": 0.0002092872506182175, "clip_ratio/region_mean": 0.001785061388121297, "completions/clipped_ratio": 0.0, "completions/max_length": 642.7, "completions/max_terminated_length": 642.7, "completions/mean_length": 197.9351608276367, "completions/mean_terminated_length": 197.9351608276367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02824337218577663, "epoch": 34.53504744413836, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.00921630859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5776461712.0, "reward": 7.907217311859131, "reward_std": 0.2855675481259823, "rewards/combined_reward/mean": 7.907216882705688, "rewards/combined_reward/std": 2.3945647716522216, "step": 56430 }, { "clip_ratio/high_max": 0.0017151413980172948, "clip_ratio/high_mean": 0.0010016091149736895, "clip_ratio/low_mean": 0.0007292465309243804, "clip_ratio/low_min": 0.00014803305578728516, "clip_ratio/region_mean": 0.0017308556280719737, "completions/clipped_ratio": 0.0, "completions/max_length": 591.1, "completions/max_terminated_length": 591.1, "completions/mean_length": 199.76797180175782, "completions/mean_terminated_length": 199.76797180175782, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.02436620166214804, "epoch": 34.55341291704928, "frac_reward_zero_std": 0.1916666731238365, "grad_norm": 0.0196533203125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5779528981.0, "reward": 8.341853427886964, "reward_std": 0.2925041504204273, "rewards/combined_reward/mean": 8.341852760314941, "rewards/combined_reward/std": 2.046210730075836, "step": 56460 }, { "clip_ratio/high_max": 0.001948762911100251, "clip_ratio/high_mean": 0.0010532787107270754, "clip_ratio/low_mean": 0.000764163356628463, "clip_ratio/low_min": 0.00019925137870207738, "clip_ratio/region_mean": 0.001817442074631496, "completions/clipped_ratio": 0.0, "completions/max_length": 568.8, "completions/max_terminated_length": 568.8, "completions/mean_length": 207.0666732788086, "completions/mean_terminated_length": 207.0666732788086, "completions/min_length": 54.6, "completions/min_terminated_length": 54.6, "entropy": 0.031039384317894776, "epoch": 34.57177838996021, "frac_reward_zero_std": 0.1375000026077032, "grad_norm": 0.01348876953125, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5782624277.0, "reward": 7.729620313644409, "reward_std": 0.31749424189329145, "rewards/combined_reward/mean": 7.729620170593262, "rewards/combined_reward/std": 2.447002184391022, "step": 56490 }, { "clip_ratio/high_max": 0.0017924381427292246, "clip_ratio/high_mean": 0.0010813563783206822, "clip_ratio/low_mean": 0.0009526098176138476, "clip_ratio/low_min": 0.00032549214083701374, "clip_ratio/region_mean": 0.0020339661956919977, "completions/clipped_ratio": 0.0, "completions/max_length": 597.1, "completions/max_terminated_length": 597.1, "completions/mean_length": 218.4627655029297, "completions/mean_terminated_length": 218.4627655029297, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.02830668807340165, "epoch": 34.59014386287114, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.020751953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5785763334.0, "reward": 7.938415956497193, "reward_std": 0.3159907966852188, "rewards/combined_reward/mean": 7.938415479660034, "rewards/combined_reward/std": 2.181919264793396, "step": 56520 }, { "clip_ratio/high_max": 0.0019172569329384715, "clip_ratio/high_mean": 0.0010874109185048534, "clip_ratio/low_mean": 0.0006870776253587489, "clip_ratio/low_min": 0.00015771340379918305, "clip_ratio/region_mean": 0.001774488547622847, "completions/clipped_ratio": 0.0, "completions/max_length": 527.2, "completions/max_terminated_length": 527.2, "completions/mean_length": 202.06745147705078, "completions/mean_terminated_length": 202.06745147705078, "completions/min_length": 40.8, "completions/min_terminated_length": 40.8, "entropy": 0.026469914157254002, "epoch": 34.60850933578206, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.00469970703125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5788839433.0, "reward": 8.201455020904541, "reward_std": 0.329569099843502, "rewards/combined_reward/mean": 8.201454687118531, "rewards/combined_reward/std": 2.0512861728668215, "step": 56550 }, { "clip_ratio/high_max": 0.0018269841117823187, "clip_ratio/high_mean": 0.0010282692222972401, "clip_ratio/low_mean": 0.0006614511599157898, "clip_ratio/low_min": 0.00012642714306518126, "clip_ratio/region_mean": 0.0016897203963405143, "completions/clipped_ratio": 0.0, "completions/max_length": 561.7, "completions/max_terminated_length": 561.7, "completions/mean_length": 206.6276077270508, "completions/mean_terminated_length": 206.6276077270508, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.028322555730119346, "epoch": 34.62687480869299, "frac_reward_zero_std": 0.13333333544433118, "grad_norm": 0.01190185546875, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5791933043.0, "reward": 8.219212961196899, "reward_std": 0.2777692168951035, "rewards/combined_reward/mean": 8.219212627410888, "rewards/combined_reward/std": 2.1296456456184387, "step": 56580 }, { "clip_ratio/high_max": 0.0020708513960319883, "clip_ratio/high_mean": 0.0010946852429090843, "clip_ratio/low_mean": 0.0008535254118517817, "clip_ratio/low_min": 0.0002899382314353716, "clip_ratio/region_mean": 0.001948210673193292, "completions/clipped_ratio": 0.0, "completions/max_length": 617.4, "completions/max_terminated_length": 617.4, "completions/mean_length": 209.36485137939454, "completions/mean_terminated_length": 209.36485137939454, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03160716813678543, "epoch": 34.64524028160392, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5795037164.0, "reward": 8.062967824935914, "reward_std": 0.34806838631629944, "rewards/combined_reward/mean": 8.062967586517335, "rewards/combined_reward/std": 2.1937616229057313, "step": 56610 }, { "clip_ratio/high_max": 0.0018796346252202057, "clip_ratio/high_mean": 0.001050256779611421, "clip_ratio/low_mean": 0.000717261959289317, "clip_ratio/low_min": 0.00021555831626756116, "clip_ratio/region_mean": 0.0017675187399921318, "completions/clipped_ratio": 0.0, "completions/max_length": 531.4, "completions/max_terminated_length": 531.4, "completions/mean_length": 203.26641235351562, "completions/mean_terminated_length": 203.26641235351562, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.028249853818366925, "epoch": 34.66360575451485, "frac_reward_zero_std": 0.2000000063329935, "grad_norm": 0.0255126953125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5798117867.0, "reward": 7.99842529296875, "reward_std": 0.28574534207582475, "rewards/combined_reward/mean": 7.998424911499024, "rewards/combined_reward/std": 2.2518448412418364, "step": 56640 }, { "clip_ratio/high_max": 0.0018918899232327627, "clip_ratio/high_mean": 0.001094144357314993, "clip_ratio/low_mean": 0.0006211017583458064, "clip_ratio/low_min": 9.62941473214111e-05, "clip_ratio/region_mean": 0.001715246119905108, "completions/clipped_ratio": 0.0, "completions/max_length": 539.9, "completions/max_terminated_length": 539.9, "completions/mean_length": 193.60625457763672, "completions/mean_terminated_length": 193.60625457763672, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029197080852463843, "epoch": 34.68197122742577, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5801161475.0, "reward": 8.069737195968628, "reward_std": 0.3692378357052803, "rewards/combined_reward/mean": 8.069736766815186, "rewards/combined_reward/std": 2.149522912502289, "step": 56670 }, { "clip_ratio/high_max": 0.001965499327828487, "clip_ratio/high_mean": 0.0010282139827419692, "clip_ratio/low_mean": 0.0007756787361965204, "clip_ratio/low_min": 0.0003020847412699368, "clip_ratio/region_mean": 0.0018038927131177236, "completions/clipped_ratio": 0.0, "completions/max_length": 523.1, "completions/max_terminated_length": 523.1, "completions/mean_length": 205.80182800292968, "completions/mean_terminated_length": 205.80182800292968, "completions/min_length": 45.3, "completions/min_terminated_length": 45.3, "entropy": 0.027922591023767988, "epoch": 34.7003367003367, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.00823974609375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5804251914.0, "reward": 8.045597553253174, "reward_std": 0.26057695895433425, "rewards/combined_reward/mean": 8.045597267150878, "rewards/combined_reward/std": 2.112693738937378, "step": 56700 }, { "clip_ratio/high_max": 0.0020890453403505185, "clip_ratio/high_mean": 0.0011745838443554628, "clip_ratio/low_mean": 0.0008253243548097088, "clip_ratio/low_min": 0.00023273036592096711, "clip_ratio/region_mean": 0.001999908206926193, "completions/clipped_ratio": 0.0, "completions/max_length": 656.1, "completions/max_terminated_length": 656.1, "completions/mean_length": 217.0812530517578, "completions/mean_terminated_length": 217.0812530517578, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.03063721931539476, "epoch": 34.71870217324763, "frac_reward_zero_std": 0.1375000037252903, "grad_norm": 0.0128173828125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5807385666.0, "reward": 8.310319423675537, "reward_std": 0.30534840524196627, "rewards/combined_reward/mean": 8.31031904220581, "rewards/combined_reward/std": 2.036524474620819, "step": 56730 }, { "clip_ratio/high_max": 0.0020922473641500497, "clip_ratio/high_mean": 0.0011005108785563304, "clip_ratio/low_mean": 0.0008714957746027115, "clip_ratio/low_min": 0.00030664315563626586, "clip_ratio/region_mean": 0.0019720066513400523, "completions/clipped_ratio": 0.0, "completions/max_length": 562.9, "completions/max_terminated_length": 562.9, "completions/mean_length": 209.83021545410156, "completions/mean_terminated_length": 209.83021545410156, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.029921881776923936, "epoch": 34.73706764615856, "frac_reward_zero_std": 0.17916667312383652, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 5810491574.0, "reward": 7.968533277511597, "reward_std": 0.3521436363458633, "rewards/combined_reward/mean": 7.96853289604187, "rewards/combined_reward/std": 2.537540888786316, "step": 56760 }, { "clip_ratio/high_max": 0.0017851536875241436, "clip_ratio/high_mean": 0.0009777872214423648, "clip_ratio/low_mean": 0.0007972843549699367, "clip_ratio/low_min": 0.00026575947704259305, "clip_ratio/region_mean": 0.0017750715701064716, "completions/clipped_ratio": 0.0, "completions/max_length": 564.4, "completions/max_terminated_length": 564.4, "completions/mean_length": 202.76016235351562, "completions/mean_terminated_length": 202.76016235351562, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.02606560042283187, "epoch": 34.75543311906948, "frac_reward_zero_std": 0.12916667200624943, "grad_norm": 0.0255126953125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5813570333.0, "reward": 8.173770475387574, "reward_std": 0.32292928248643876, "rewards/combined_reward/mean": 8.173770236968995, "rewards/combined_reward/std": 2.0161997735500337, "step": 56790 }, { "clip_ratio/high_max": 0.0021776962001846794, "clip_ratio/high_mean": 0.00119665852447118, "clip_ratio/low_mean": 0.0006987736473092809, "clip_ratio/low_min": 0.0002639245004199135, "clip_ratio/region_mean": 0.0018954321455870134, "completions/clipped_ratio": 0.0, "completions/max_length": 527.2, "completions/max_terminated_length": 527.2, "completions/mean_length": 210.74063110351562, "completions/mean_terminated_length": 210.74063110351562, "completions/min_length": 43.6, "completions/min_terminated_length": 43.6, "entropy": 0.028297805258383355, "epoch": 34.77379859198041, "frac_reward_zero_std": 0.16250000558793545, "grad_norm": 0.0179443359375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5816679737.0, "reward": 8.153527116775512, "reward_std": 0.2922630086541176, "rewards/combined_reward/mean": 8.153527021408081, "rewards/combined_reward/std": 2.257323145866394, "step": 56820 }, { "clip_ratio/high_max": 0.0018051912632169357, "clip_ratio/high_mean": 0.0009891508611569104, "clip_ratio/low_mean": 0.0007242346132746509, "clip_ratio/low_min": 0.0001862826010134692, "clip_ratio/region_mean": 0.0017133854770994125, "completions/clipped_ratio": 0.0, "completions/max_length": 508.3, "completions/max_terminated_length": 508.3, "completions/mean_length": 207.69271240234374, "completions/mean_terminated_length": 207.69271240234374, "completions/min_length": 49.3, "completions/min_terminated_length": 49.3, "entropy": 0.027734982129186393, "epoch": 34.79216406489134, "frac_reward_zero_std": 0.17083333916962146, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5819777437.0, "reward": 8.157039976119995, "reward_std": 0.2494307592511177, "rewards/combined_reward/mean": 8.15703992843628, "rewards/combined_reward/std": 2.0908753871917725, "step": 56850 }, { "clip_ratio/high_max": 0.0022543122623270998, "clip_ratio/high_mean": 0.0013131544966502892, "clip_ratio/low_mean": 0.000817711098958777, "clip_ratio/low_min": 0.00018186946278243946, "clip_ratio/region_mean": 0.002130865605431609, "completions/clipped_ratio": 0.0, "completions/max_length": 653.5, "completions/max_terminated_length": 653.5, "completions/mean_length": 201.3572982788086, "completions/mean_terminated_length": 201.3572982788086, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.03265062068433811, "epoch": 34.81052953780227, "frac_reward_zero_std": 0.2083333410322666, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5822850809.0, "reward": 7.9614335060119625, "reward_std": 0.266403903067112, "rewards/combined_reward/mean": 7.961433172225952, "rewards/combined_reward/std": 2.2600341558456423, "step": 56880 }, { "clip_ratio/high_max": 0.002250987033282096, "clip_ratio/high_mean": 0.0012542735700359724, "clip_ratio/low_mean": 0.0008203023964597378, "clip_ratio/low_min": 0.00023948502105971178, "clip_ratio/region_mean": 0.002074575956794433, "completions/clipped_ratio": 0.0, "completions/max_length": 576.6, "completions/max_terminated_length": 576.6, "completions/mean_length": 203.4612060546875, "completions/mean_terminated_length": 203.4612060546875, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.027611014774690072, "epoch": 34.82889501071319, "frac_reward_zero_std": 0.21250000819563866, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 5825932260.0, "reward": 8.14414873123169, "reward_std": 0.289215287566185, "rewards/combined_reward/mean": 8.144148540496825, "rewards/combined_reward/std": 2.339837431907654, "step": 56910 }, { "clip_ratio/high_max": 0.0030367423760859916, "clip_ratio/high_mean": 0.0014884739818323094, "clip_ratio/low_mean": 0.000744040798599599, "clip_ratio/low_min": 0.00018837866227841004, "clip_ratio/region_mean": 0.0022325147777640573, "completions/clipped_ratio": 0.0, "completions/max_length": 596.2, "completions/max_terminated_length": 596.2, "completions/mean_length": 201.67396240234376, "completions/mean_terminated_length": 201.67396240234376, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "entropy": 0.03294335240498185, "epoch": 34.84726048362412, "frac_reward_zero_std": 0.18333333916962147, "grad_norm": 0.01458740234375, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 5829006848.0, "reward": 8.264650058746337, "reward_std": 0.3257413744926453, "rewards/combined_reward/mean": 8.264649868011475, "rewards/combined_reward/std": 2.065820908546448, "step": 56940 }, { "clip_ratio/high_max": 0.0017707323238331204, "clip_ratio/high_mean": 0.0009814575316947109, "clip_ratio/low_mean": 0.0006550794232074016, "clip_ratio/low_min": 0.00015434899153963973, "clip_ratio/region_mean": 0.0016365369699390916, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 204.65521545410155, "completions/mean_terminated_length": 204.65521545410155, "completions/min_length": 53.4, "completions/min_terminated_length": 53.4, "entropy": 0.028722061961889266, "epoch": 34.86562595653505, "frac_reward_zero_std": 0.1583333384245634, "grad_norm": 0.0277099609375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5832092884.0, "reward": 8.266582775115968, "reward_std": 0.304014128446579, "rewards/combined_reward/mean": 8.266582727432251, "rewards/combined_reward/std": 1.9537837982177735, "step": 56970 }, { "clip_ratio/high_max": 0.0019433374186822523, "clip_ratio/high_mean": 0.0011547278091408467, "clip_ratio/low_mean": 0.000831237106952661, "clip_ratio/low_min": 0.00026833892212986635, "clip_ratio/region_mean": 0.001985964902754252, "completions/clipped_ratio": 0.0, "completions/max_length": 627.4, "completions/max_terminated_length": 627.4, "completions/mean_length": 210.74740142822264, "completions/mean_terminated_length": 210.74740142822264, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.029067234632869563, "epoch": 34.88399142944598, "frac_reward_zero_std": 0.17500000335276128, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.0028, "num_tokens": 5835202314.0, "reward": 8.261504745483398, "reward_std": 0.3837428867816925, "rewards/combined_reward/mean": 8.261504554748536, "rewards/combined_reward/std": 2.069579875469208, "step": 57000 }, { "clip_ratio/high_max": 0.0016472013422268597, "clip_ratio/high_mean": 0.0009728974119449656, "clip_ratio/low_mean": 0.0008225684787854941, "clip_ratio/low_min": 0.00022081519297595758, "clip_ratio/region_mean": 0.0017954658726618315, "completions/clipped_ratio": 0.0, "completions/max_length": 657.7, "completions/max_terminated_length": 657.7, "completions/mean_length": 214.61875762939454, "completions/mean_terminated_length": 214.61875762939454, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.027153771044686435, "epoch": 34.9023569023569, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0244140625, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 5838326610.0, "reward": 7.977551221847534, "reward_std": 0.3590838462114334, "rewards/combined_reward/mean": 7.977550792694092, "rewards/combined_reward/std": 2.146362233161926, "step": 57030 }, { "clip_ratio/high_max": 0.0022244515843340197, "clip_ratio/high_mean": 0.001168477738732084, "clip_ratio/low_mean": 0.0007999684081975526, "clip_ratio/low_min": 0.00017046852638789763, "clip_ratio/region_mean": 0.0019684461445043174, "completions/clipped_ratio": 0.0, "completions/max_length": 555.9, "completions/max_terminated_length": 555.9, "completions/mean_length": 189.36146697998046, "completions/mean_terminated_length": 189.36146697998046, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.0320206206291914, "epoch": 34.92072237526783, "frac_reward_zero_std": 0.1666666716337204, "grad_norm": 0.0245361328125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5841353918.0, "reward": 8.309766244888305, "reward_std": 0.30450617522001266, "rewards/combined_reward/mean": 8.309765672683715, "rewards/combined_reward/std": 1.863964307308197, "step": 57060 }, { "clip_ratio/high_max": 0.002000160384341143, "clip_ratio/high_mean": 0.001174150122339294, "clip_ratio/low_mean": 0.0007520202991145198, "clip_ratio/low_min": 0.00021329785134488095, "clip_ratio/region_mean": 0.001926170423394069, "completions/clipped_ratio": 0.0, "completions/max_length": 619.5, "completions/max_terminated_length": 619.5, "completions/mean_length": 203.59740295410157, "completions/mean_terminated_length": 203.59740295410157, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.029010262355829278, "epoch": 34.93908784817876, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.01324462890625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5844435892.0, "reward": 7.935522794723511, "reward_std": 0.24639570713043213, "rewards/combined_reward/mean": 7.935522651672363, "rewards/combined_reward/std": 2.5637630939483644, "step": 57090 }, { "clip_ratio/high_max": 0.0017747920021065511, "clip_ratio/high_mean": 0.0010233091367505647, "clip_ratio/low_mean": 0.0007958265952765942, "clip_ratio/low_min": 0.000290175241389079, "clip_ratio/region_mean": 0.001819135722568414, "completions/clipped_ratio": 0.0, "completions/max_length": 554.2, "completions/max_terminated_length": 554.2, "completions/mean_length": 205.0041717529297, "completions/mean_terminated_length": 205.0041717529297, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.02708816167432815, "epoch": 34.95745332108969, "frac_reward_zero_std": 0.18333333730697632, "grad_norm": 0.006317138671875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 5847523268.0, "reward": 7.945837450027466, "reward_std": 0.23591796457767486, "rewards/combined_reward/mean": 7.945837211608887, "rewards/combined_reward/std": 2.308567988872528, "step": 57120 }, { "clip_ratio/high_max": 0.0020095436736786115, "clip_ratio/high_mean": 0.0012009284218947869, "clip_ratio/low_mean": 0.000518548834467462, "clip_ratio/low_min": 0.00010824616086514046, "clip_ratio/region_mean": 0.0017194772517541423, "completions/clipped_ratio": 0.0, "completions/max_length": 635.4, "completions/max_terminated_length": 635.4, "completions/mean_length": 199.88333892822266, "completions/mean_terminated_length": 199.88333892822266, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.02645539479951064, "epoch": 34.97581879400061, "frac_reward_zero_std": 0.1958333395421505, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5850590980.0, "reward": 8.092255306243896, "reward_std": 0.3179690703749657, "rewards/combined_reward/mean": 8.092254972457885, "rewards/combined_reward/std": 2.258890354633331, "step": 57150 }, { "clip_ratio/high_max": 0.0019507629597986427, "clip_ratio/high_mean": 0.0011646075700506723, "clip_ratio/low_mean": 0.000708529791154433, "clip_ratio/low_min": 0.00020053309126524254, "clip_ratio/region_mean": 0.0018731373575671265, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 201.90911712646485, "completions/mean_terminated_length": 201.90911712646485, "completions/min_length": 45.3, "completions/min_terminated_length": 45.3, "entropy": 0.026149231443802514, "epoch": 34.99418426691154, "frac_reward_zero_std": 0.16666667051613332, "grad_norm": 0.00732421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5853666471.0, "reward": 8.107760047912597, "reward_std": 0.296064256131649, "rewards/combined_reward/mean": 8.107759952545166, "rewards/combined_reward/std": 2.2071680426597595, "step": 57180 }, { "clip_ratio/high_max": 0.0023034514509118545, "clip_ratio/high_mean": 0.0012552310414679088, "clip_ratio/low_mean": 0.0009779132459862311, "clip_ratio/low_min": 0.00022836126097035213, "clip_ratio/region_mean": 0.0022331442541755797, "completions/clipped_ratio": 0.0, "completions/max_length": 528.1, "completions/max_terminated_length": 528.1, "completions/mean_length": 207.39557952880858, "completions/mean_terminated_length": 207.39557952880858, "completions/min_length": 44.4, "completions/min_terminated_length": 44.4, "entropy": 0.02816517185419798, "epoch": 35.01285583103765, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.0079345703125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5856763030.0, "reward": 8.194358348846436, "reward_std": 0.3204363971948624, "rewards/combined_reward/mean": 8.194358110427856, "rewards/combined_reward/std": 2.1526938140392304, "step": 57210 }, { "clip_ratio/high_max": 0.0025799535714516727, "clip_ratio/high_mean": 0.0014360884111132084, "clip_ratio/low_mean": 0.000991759420260981, "clip_ratio/low_min": 0.0003295440939837135, "clip_ratio/region_mean": 0.0024278478051807423, "completions/clipped_ratio": 0.0, "completions/max_length": 514.9, "completions/max_terminated_length": 514.9, "completions/mean_length": 197.0109405517578, "completions/mean_terminated_length": 197.0109405517578, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.028791628079488873, "epoch": 35.03122130394858, "frac_reward_zero_std": 0.16666666828095913, "grad_norm": 0.027099609375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5859819712.0, "reward": 7.8776264667510985, "reward_std": 0.30081100612878797, "rewards/combined_reward/mean": 7.877626371383667, "rewards/combined_reward/std": 2.440717315673828, "step": 57240 }, { "clip_ratio/high_max": 0.002428742448197833, "clip_ratio/high_mean": 0.0015043075135812009, "clip_ratio/low_mean": 0.0010884106735223515, "clip_ratio/low_min": 0.0002554054228918782, "clip_ratio/region_mean": 0.0025927181898926697, "completions/clipped_ratio": 0.0, "completions/max_length": 561.9, "completions/max_terminated_length": 561.9, "completions/mean_length": 211.39974365234374, "completions/mean_terminated_length": 211.39974365234374, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.03292849484520654, "epoch": 35.049586776859506, "frac_reward_zero_std": 0.1750000026077032, "grad_norm": 0.0106201171875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5862931647.0, "reward": 8.239630937576294, "reward_std": 0.3078101083636284, "rewards/combined_reward/mean": 8.239630603790284, "rewards/combined_reward/std": 1.9554160594940186, "step": 57270 }, { "clip_ratio/high_max": 0.002083800492497782, "clip_ratio/high_mean": 0.0012438321466712902, "clip_ratio/low_mean": 0.0008959031187259825, "clip_ratio/low_min": 0.00024209978631309543, "clip_ratio/region_mean": 0.0021397352638208153, "completions/clipped_ratio": 0.0, "completions/max_length": 505.6, "completions/max_terminated_length": 505.6, "completions/mean_length": 203.59062805175782, "completions/mean_terminated_length": 203.59062805175782, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.025072975154034793, "epoch": 35.06795224977043, "frac_reward_zero_std": 0.12083333656191826, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5866013595.0, "reward": 8.299467039108276, "reward_std": 0.3065675422549248, "rewards/combined_reward/mean": 8.299466514587403, "rewards/combined_reward/std": 2.1014126777648925, "step": 57300 }, { "clip_ratio/high_max": 0.002485567306333299, "clip_ratio/high_mean": 0.0013618374539873912, "clip_ratio/low_mean": 0.0010449251193980066, "clip_ratio/low_min": 0.0002887007479633515, "clip_ratio/region_mean": 0.002406762563138424, "completions/clipped_ratio": 0.0, "completions/max_length": 604.1, "completions/max_terminated_length": 604.1, "completions/mean_length": 201.38880767822266, "completions/mean_terminated_length": 201.38880767822266, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "entropy": 0.029670341530193885, "epoch": 35.08631772268136, "frac_reward_zero_std": 0.15833333805203437, "grad_norm": 0.0076904296875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5869087088.0, "reward": 8.126554822921753, "reward_std": 0.27224249541759493, "rewards/combined_reward/mean": 8.126554822921753, "rewards/combined_reward/std": 2.104246956110001, "step": 57330 }, { "clip_ratio/high_max": 0.0024155464525392745, "clip_ratio/high_mean": 0.0014100435444561299, "clip_ratio/low_mean": 0.0009176156510269114, "clip_ratio/low_min": 0.0003272146881499793, "clip_ratio/region_mean": 0.0023276591911174666, "completions/clipped_ratio": 0.0, "completions/max_length": 614.5, "completions/max_terminated_length": 614.5, "completions/mean_length": 205.1705795288086, "completions/mean_terminated_length": 205.1705795288086, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02762652236657838, "epoch": 35.10468319559229, "frac_reward_zero_std": 0.1875000052154064, "grad_norm": 0.0069580078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5872175103.0, "reward": 8.057958984375, "reward_std": 0.33029039204120636, "rewards/combined_reward/mean": 8.057958793640136, "rewards/combined_reward/std": 2.072069025039673, "step": 57360 }, { "clip_ratio/high_max": 0.0021193441677799757, "clip_ratio/high_mean": 0.0011736061054155773, "clip_ratio/low_mean": 0.000999212848788981, "clip_ratio/low_min": 0.0002663086076305869, "clip_ratio/region_mean": 0.002172818943411888, "completions/clipped_ratio": 0.0, "completions/max_length": 679.8, "completions/max_terminated_length": 679.8, "completions/mean_length": 206.64818267822267, "completions/mean_terminated_length": 206.64818267822267, "completions/min_length": 48.8, "completions/min_terminated_length": 48.8, "entropy": 0.02800765836145729, "epoch": 35.123048668503216, "frac_reward_zero_std": 0.1333333358168602, "grad_norm": 0.013427734375, "learning_rate": 5e-06, "loss": 0.0024, "num_tokens": 5875268792.0, "reward": 7.961542701721191, "reward_std": 0.295182553678751, "rewards/combined_reward/mean": 7.961542415618896, "rewards/combined_reward/std": 2.5251262068748472, "step": 57390 }, { "clip_ratio/high_max": 0.002688046459419032, "clip_ratio/high_mean": 0.0014937183371027155, "clip_ratio/low_mean": 0.0009405692157694527, "clip_ratio/low_min": 0.00029740999816567636, "clip_ratio/region_mean": 0.0024342875253447953, "completions/clipped_ratio": 0.0, "completions/max_length": 593.4, "completions/max_terminated_length": 593.4, "completions/mean_length": 194.93828582763672, "completions/mean_terminated_length": 194.93828582763672, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.029243756163244445, "epoch": 35.14141414141414, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.025390625, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5878317515.0, "reward": 7.8506029605865475, "reward_std": 0.32491273134946824, "rewards/combined_reward/mean": 7.850602865219116, "rewards/combined_reward/std": 2.423762547969818, "step": 57420 }, { "clip_ratio/high_max": 0.0027007840161483426, "clip_ratio/high_mean": 0.0014781786247719233, "clip_ratio/low_mean": 0.0010244005594358895, "clip_ratio/low_min": 0.00031648785904205093, "clip_ratio/region_mean": 0.0025025791799028712, "completions/clipped_ratio": 0.0, "completions/max_length": 524.4, "completions/max_terminated_length": 524.4, "completions/mean_length": 204.9953170776367, "completions/mean_terminated_length": 204.9953170776367, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.028336535564934213, "epoch": 35.15977961432507, "frac_reward_zero_std": 0.14583333656191827, "grad_norm": 0.0081787109375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5881404857.0, "reward": 8.194913768768311, "reward_std": 0.3033048823475838, "rewards/combined_reward/mean": 8.194913625717163, "rewards/combined_reward/std": 2.080424028635025, "step": 57450 }, { "clip_ratio/high_max": 0.002287738534505479, "clip_ratio/high_mean": 0.0012358835208336435, "clip_ratio/low_mean": 0.0009243439424608368, "clip_ratio/low_min": 0.0002546793487757289, "clip_ratio/region_mean": 0.002160227452259278, "completions/clipped_ratio": 0.0, "completions/max_length": 602.1, "completions/max_terminated_length": 602.1, "completions/mean_length": 207.28907012939453, "completions/mean_terminated_length": 207.28907012939453, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.025119458410578468, "epoch": 35.178145087236, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.009033203125, "learning_rate": 5e-06, "loss": 0.0002, "num_tokens": 5884501007.0, "reward": 7.983792638778686, "reward_std": 0.2886705592274666, "rewards/combined_reward/mean": 7.983792352676391, "rewards/combined_reward/std": 2.4269787073135376, "step": 57480 }, { "clip_ratio/high_max": 0.002493076269456651, "clip_ratio/high_mean": 0.001449880294482379, "clip_ratio/low_mean": 0.0009466470015468076, "clip_ratio/low_min": 0.00018097657569645284, "clip_ratio/region_mean": 0.0023965273178570595, "completions/clipped_ratio": 0.0, "completions/max_length": 522.4, "completions/max_terminated_length": 522.4, "completions/mean_length": 191.79114990234376, "completions/mean_terminated_length": 191.79114990234376, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02734696646220982, "epoch": 35.196510560146926, "frac_reward_zero_std": 0.20416667349636555, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5887537645.0, "reward": 8.11837477684021, "reward_std": 0.2708499327301979, "rewards/combined_reward/mean": 8.118374490737915, "rewards/combined_reward/std": 2.1960694193840027, "step": 57510 }, { "clip_ratio/high_max": 0.002589409574769282, "clip_ratio/high_mean": 0.001565620073476263, "clip_ratio/low_mean": 0.0007380320418936511, "clip_ratio/low_min": 0.00015646145426823448, "clip_ratio/region_mean": 0.0023036521277390422, "completions/clipped_ratio": 0.0, "completions/max_length": 629.8, "completions/max_terminated_length": 629.8, "completions/mean_length": 203.45260925292968, "completions/mean_terminated_length": 203.45260925292968, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.02748859195659558, "epoch": 35.21487603305785, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.0177001953125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5890619063.0, "reward": 8.054615497589111, "reward_std": 0.29468039721250533, "rewards/combined_reward/mean": 8.054615211486816, "rewards/combined_reward/std": 2.4037843108177186, "step": 57540 }, { "clip_ratio/high_max": 0.002853063454676885, "clip_ratio/high_mean": 0.001568459247694894, "clip_ratio/low_mean": 0.0009648893948906335, "clip_ratio/low_min": 0.00021802430055686272, "clip_ratio/region_mean": 0.0025333486138454948, "completions/clipped_ratio": 0.0, "completions/max_length": 607.5, "completions/max_terminated_length": 607.5, "completions/mean_length": 197.5768280029297, "completions/mean_terminated_length": 197.5768280029297, "completions/min_length": 38.5, "completions/min_terminated_length": 38.5, "entropy": 0.03285857943507532, "epoch": 35.23324150596878, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.0033721923828125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 5893677918.0, "reward": 7.99119725227356, "reward_std": 0.3228192120790482, "rewards/combined_reward/mean": 7.991196918487549, "rewards/combined_reward/std": 2.236416959762573, "step": 57570 }, { "clip_ratio/high_max": 0.002285773714296132, "clip_ratio/high_mean": 0.00130519061403902, "clip_ratio/low_mean": 0.0007526689112031212, "clip_ratio/low_min": 0.00016978654651514564, "clip_ratio/region_mean": 0.002057859510144529, "completions/clipped_ratio": 0.0, "completions/max_length": 667.6, "completions/max_terminated_length": 667.6, "completions/mean_length": 203.96224670410157, "completions/mean_terminated_length": 203.96224670410157, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.02505368129350245, "epoch": 35.25160697887971, "frac_reward_zero_std": 0.15833334028720855, "grad_norm": 0.032470703125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5896761293.0, "reward": 8.19630126953125, "reward_std": 0.3184932589530945, "rewards/combined_reward/mean": 8.196301174163818, "rewards/combined_reward/std": 2.143882143497467, "step": 57600 }, { "clip_ratio/high_max": 0.002612294193143801, "clip_ratio/high_mean": 0.0015223307408935701, "clip_ratio/low_mean": 0.001040084810423044, "clip_ratio/low_min": 0.000290428650866185, "clip_ratio/region_mean": 0.002562415572659423, "completions/clipped_ratio": 0.0, "completions/max_length": 611.9, "completions/max_terminated_length": 611.9, "completions/mean_length": 194.30938110351562, "completions/mean_terminated_length": 194.30938110351562, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.02948221635694305, "epoch": 35.269972451790636, "frac_reward_zero_std": 0.18750000596046448, "grad_norm": 0.038330078125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5899807601.0, "reward": 8.225948238372803, "reward_std": 0.2525905042886734, "rewards/combined_reward/mean": 8.225947856903076, "rewards/combined_reward/std": 2.0249802350997923, "step": 57630 }, { "clip_ratio/high_max": 0.0022406095168359267, "clip_ratio/high_mean": 0.0012335766080165437, "clip_ratio/low_mean": 0.0009024180869649475, "clip_ratio/low_min": 0.00024217233221861533, "clip_ratio/region_mean": 0.00213599469861947, "completions/clipped_ratio": 0.0, "completions/max_length": 681.5, "completions/max_terminated_length": 681.5, "completions/mean_length": 195.87239990234374, "completions/mean_terminated_length": 195.87239990234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02669737102308621, "epoch": 35.28833792470156, "frac_reward_zero_std": 0.22500000819563865, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5902859911.0, "reward": 7.9086230278015135, "reward_std": 0.2922117725014687, "rewards/combined_reward/mean": 7.9086226463317875, "rewards/combined_reward/std": 2.5181735277175905, "step": 57660 }, { "clip_ratio/high_max": 0.002236807041723902, "clip_ratio/high_mean": 0.0013166519118385622, "clip_ratio/low_mean": 0.0009556119767997491, "clip_ratio/low_min": 0.00021194189175730572, "clip_ratio/region_mean": 0.00227226386826563, "completions/clipped_ratio": 0.0, "completions/max_length": 578.7, "completions/max_terminated_length": 578.7, "completions/mean_length": 199.21537170410156, "completions/mean_terminated_length": 199.21537170410156, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.026225357176736, "epoch": 35.30670339761249, "frac_reward_zero_std": 0.17500000447034836, "grad_norm": 0.006805419921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5905925058.0, "reward": 8.04232416152954, "reward_std": 0.36811877936124804, "rewards/combined_reward/mean": 8.04232416152954, "rewards/combined_reward/std": 2.142400121688843, "step": 57690 }, { "clip_ratio/high_max": 0.0023761111282510684, "clip_ratio/high_mean": 0.001420256572115856, "clip_ratio/low_mean": 0.0009161144174261912, "clip_ratio/low_min": 0.00026035876605116454, "clip_ratio/region_mean": 0.0023363709779611478, "completions/clipped_ratio": 0.0, "completions/max_length": 590.2, "completions/max_terminated_length": 590.2, "completions/mean_length": 205.1994842529297, "completions/mean_terminated_length": 205.1994842529297, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.028981383393208184, "epoch": 35.32506887052342, "frac_reward_zero_std": 0.12916667088866235, "grad_norm": 0.006866455078125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5909013184.0, "reward": 8.279846668243408, "reward_std": 0.35741655379533765, "rewards/combined_reward/mean": 8.279846477508546, "rewards/combined_reward/std": 2.0297155737876893, "step": 57720 }, { "clip_ratio/high_max": 0.0023817904007349473, "clip_ratio/high_mean": 0.0013565181592033089, "clip_ratio/low_mean": 0.0011597937662979044, "clip_ratio/low_min": 0.00043322549245203843, "clip_ratio/region_mean": 0.002516311927077671, "completions/clipped_ratio": 0.0, "completions/max_length": 629.7, "completions/max_terminated_length": 629.7, "completions/mean_length": 205.68437957763672, "completions/mean_terminated_length": 205.68437957763672, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.030899282870814206, "epoch": 35.343434343434346, "frac_reward_zero_std": 0.19166667386889458, "grad_norm": 0.01708984375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5912103172.0, "reward": 7.839525842666626, "reward_std": 0.3249225154519081, "rewards/combined_reward/mean": 7.839525318145752, "rewards/combined_reward/std": 2.6443165063858034, "step": 57750 }, { "clip_ratio/high_max": 0.0025779447678360157, "clip_ratio/high_mean": 0.0014385433006585421, "clip_ratio/low_mean": 0.0008878108600280636, "clip_ratio/low_min": 0.00024045024765655397, "clip_ratio/region_mean": 0.002326354158867616, "completions/clipped_ratio": 0.0, "completions/max_length": 588.8, "completions/max_terminated_length": 588.8, "completions/mean_length": 202.8140670776367, "completions/mean_terminated_length": 202.8140670776367, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.028649656008929013, "epoch": 35.36179981634527, "frac_reward_zero_std": 0.15000000447034836, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5915182138.0, "reward": 8.328951978683472, "reward_std": 0.29218452125787736, "rewards/combined_reward/mean": 8.328951692581176, "rewards/combined_reward/std": 1.9770577073097229, "step": 57780 }, { "clip_ratio/high_max": 0.0025359957508044317, "clip_ratio/high_mean": 0.0015199996913603778, "clip_ratio/low_mean": 0.001005294937203871, "clip_ratio/low_min": 0.00024537644276279025, "clip_ratio/region_mean": 0.0025252946268058923, "completions/clipped_ratio": 0.0, "completions/max_length": 605.7, "completions/max_terminated_length": 605.7, "completions/mean_length": 204.00391082763673, "completions/mean_terminated_length": 204.00391082763673, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.029507604039584596, "epoch": 35.3801652892562, "frac_reward_zero_std": 0.1833333395421505, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5918265673.0, "reward": 7.958428859710693, "reward_std": 0.33312688767910004, "rewards/combined_reward/mean": 7.958428573608399, "rewards/combined_reward/std": 2.401227557659149, "step": 57810 }, { "clip_ratio/high_max": 0.0024288541462738067, "clip_ratio/high_mean": 0.0014407087789246968, "clip_ratio/low_mean": 0.0011788872065759885, "clip_ratio/low_min": 0.00039818076305285407, "clip_ratio/region_mean": 0.0026195959915639834, "completions/clipped_ratio": 0.0, "completions/max_length": 548.2, "completions/max_terminated_length": 548.2, "completions/mean_length": 209.96302642822266, "completions/mean_terminated_length": 209.96302642822266, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.0327973892679438, "epoch": 35.39853076216713, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0108642578125, "learning_rate": 5e-06, "loss": 0.0023, "num_tokens": 5921372091.0, "reward": 7.732763814926147, "reward_std": 0.36107536405324936, "rewards/combined_reward/mean": 7.732763576507568, "rewards/combined_reward/std": 2.5217628598213198, "step": 57840 }, { "clip_ratio/high_max": 0.002268353564431891, "clip_ratio/high_mean": 0.0011920737256635523, "clip_ratio/low_mean": 0.0008704871218166469, "clip_ratio/low_min": 0.0001442205718679664, "clip_ratio/region_mean": 0.002062560836687529, "completions/clipped_ratio": 0.0, "completions/max_length": 573.5, "completions/max_terminated_length": 573.5, "completions/mean_length": 194.86068115234374, "completions/mean_terminated_length": 194.86068115234374, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "entropy": 0.025621574899802605, "epoch": 35.416896235078056, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0021, "num_tokens": 5924420516.0, "reward": 7.858017492294311, "reward_std": 0.3560347095131874, "rewards/combined_reward/mean": 7.858017110824585, "rewards/combined_reward/std": 2.3611268281936644, "step": 57870 }, { "clip_ratio/high_max": 0.0021283454339330395, "clip_ratio/high_mean": 0.001241658953707277, "clip_ratio/low_mean": 0.000782820593061236, "clip_ratio/low_min": 0.00013699931478186044, "clip_ratio/region_mean": 0.0020244795605928326, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 206.40703582763672, "completions/mean_terminated_length": 206.40703582763672, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.025926455467318494, "epoch": 35.43526170798898, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.01287841796875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5927513279.0, "reward": 8.00201644897461, "reward_std": 0.3425240948796272, "rewards/combined_reward/mean": 8.002016305923462, "rewards/combined_reward/std": 2.5325142621994017, "step": 57900 }, { "clip_ratio/high_max": 0.0023061530574826367, "clip_ratio/high_mean": 0.0013659241568044914, "clip_ratio/low_mean": 0.0008918789987850081, "clip_ratio/low_min": 0.0002938988684036303, "clip_ratio/region_mean": 0.002257803174628255, "completions/clipped_ratio": 0.0, "completions/max_length": 643.1, "completions/max_terminated_length": 643.1, "completions/mean_length": 205.02292175292968, "completions/mean_terminated_length": 205.02292175292968, "completions/min_length": 44.8, "completions/min_terminated_length": 44.8, "entropy": 0.03055855051303903, "epoch": 35.45362718089991, "frac_reward_zero_std": 0.20416667461395263, "grad_norm": 0.015625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5930600727.0, "reward": 8.214926624298096, "reward_std": 0.3248918175697327, "rewards/combined_reward/mean": 8.214926528930665, "rewards/combined_reward/std": 2.2461347818374633, "step": 57930 }, { "clip_ratio/high_max": 0.002534433649755859, "clip_ratio/high_mean": 0.001430254298611544, "clip_ratio/low_mean": 0.0011388662929675774, "clip_ratio/low_min": 0.0003514799721112164, "clip_ratio/region_mean": 0.0025691205887900044, "completions/clipped_ratio": 0.0, "completions/max_length": 556.9, "completions/max_terminated_length": 556.9, "completions/mean_length": 199.63125457763672, "completions/mean_terminated_length": 199.63125457763672, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.03122339742258191, "epoch": 35.47199265381084, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 5933667471.0, "reward": 7.833720064163208, "reward_std": 0.3386207178235054, "rewards/combined_reward/mean": 7.8337198257446286, "rewards/combined_reward/std": 2.4288536667823792, "step": 57960 }, { "clip_ratio/high_max": 0.002375945230596699, "clip_ratio/high_mean": 0.00138067472265296, "clip_ratio/low_mean": 0.0009148355934788317, "clip_ratio/low_min": 0.0002912285510925964, "clip_ratio/region_mean": 0.0022955103345642176, "completions/clipped_ratio": 0.0, "completions/max_length": 711.7, "completions/max_terminated_length": 711.7, "completions/mean_length": 203.30833892822267, "completions/mean_terminated_length": 203.30833892822267, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028070614067837597, "epoch": 35.490358126721766, "frac_reward_zero_std": 0.16666667126119136, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 5936748335.0, "reward": 8.163441371917724, "reward_std": 0.4378213346004486, "rewards/combined_reward/mean": 8.163440895080566, "rewards/combined_reward/std": 2.0064005732536314, "step": 57990 }, { "clip_ratio/high_max": 0.0017241781264601742, "clip_ratio/high_mean": 0.000992600510289776, "clip_ratio/low_mean": 0.0006142636635559029, "clip_ratio/low_min": 0.00018092639329552184, "clip_ratio/region_mean": 0.0016068641823949292, "completions/clipped_ratio": 0.0, "completions/max_length": 633.1428571428571, "completions/max_terminated_length": 633.1428571428571, "completions/mean_length": 199.79129900251115, "completions/mean_terminated_length": 199.79129900251115, "completions/min_length": 37.714285714285715, "completions/min_terminated_length": 37.714285714285715, "entropy": 0.027648468944244088, "epoch": 35.50811141720233, "frac_reward_zero_std": 0.17261905329568045, "grad_norm": 0.01019287109375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5939825492.0, "reward": 7.8416836602347235, "reward_std": 0.3240886926651001, "rewards/combined_reward/mean": 7.841683319636753, "rewards/combined_reward/std": 2.4810519218444824, "step": 58020 }, { "clip_ratio/high_max": 0.0019198395699883501, "clip_ratio/high_mean": 0.0010944720849996278, "clip_ratio/low_mean": 0.0006858610607499335, "clip_ratio/low_min": 0.0001342104532038017, "clip_ratio/region_mean": 0.0017803331410201886, "completions/clipped_ratio": 0.0, "completions/max_length": 644.2, "completions/max_terminated_length": 644.2, "completions/mean_length": 185.82370147705078, "completions/mean_terminated_length": 185.82370147705078, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.027923740353435277, "epoch": 35.52647689011325, "frac_reward_zero_std": 0.2125000074505806, "grad_norm": 0.01104736328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5942839215.0, "reward": 7.796386003494263, "reward_std": 0.32175343334674833, "rewards/combined_reward/mean": 7.796385955810547, "rewards/combined_reward/std": 2.5720159292221068, "step": 58050 }, { "clip_ratio/high_max": 0.0017423351717297919, "clip_ratio/high_mean": 0.0009426034309096091, "clip_ratio/low_mean": 0.0008071038693742594, "clip_ratio/low_min": 0.00027211239430471326, "clip_ratio/region_mean": 0.0017497072966458896, "completions/clipped_ratio": 0.0, "completions/max_length": 589.5, "completions/max_terminated_length": 589.5, "completions/mean_length": 209.4705780029297, "completions/mean_terminated_length": 209.4705780029297, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.02786045690687994, "epoch": 35.54484236302418, "frac_reward_zero_std": 0.14583333879709243, "grad_norm": 0.0194091796875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5945943742.0, "reward": 8.253332710266113, "reward_std": 0.3083075970411301, "rewards/combined_reward/mean": 8.253332471847534, "rewards/combined_reward/std": 2.1953527092933656, "step": 58080 }, { "clip_ratio/high_max": 0.0019717786150674024, "clip_ratio/high_mean": 0.001090005259902682, "clip_ratio/low_mean": 0.0007936748015708872, "clip_ratio/low_min": 0.00021241362434617865, "clip_ratio/region_mean": 0.0018836800523179893, "completions/clipped_ratio": 0.0, "completions/max_length": 652.6, "completions/max_terminated_length": 652.6, "completions/mean_length": 199.68594207763672, "completions/mean_terminated_length": 199.68594207763672, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.030908959172666072, "epoch": 35.56320783593511, "frac_reward_zero_std": 0.2125000052154064, "grad_norm": 0.0185546875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5949010696.0, "reward": 7.9664342403411865, "reward_std": 0.25295228511095047, "rewards/combined_reward/mean": 7.966433715820313, "rewards/combined_reward/std": 2.0832414984703065, "step": 58110 }, { "clip_ratio/high_max": 0.0018816295084737552, "clip_ratio/high_mean": 0.001101654808129145, "clip_ratio/low_mean": 0.0007473396683659909, "clip_ratio/low_min": 0.00017343943278926114, "clip_ratio/region_mean": 0.0018489944864995777, "completions/clipped_ratio": 0.0, "completions/max_length": 618.1, "completions/max_terminated_length": 618.1, "completions/mean_length": 211.3658920288086, "completions/mean_terminated_length": 211.3658920288086, "completions/min_length": 52.0, "completions/min_terminated_length": 52.0, "entropy": 0.028864889731630684, "epoch": 35.58157330884604, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.01318359375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 5952122501.0, "reward": 7.9375639915466305, "reward_std": 0.3218672931194305, "rewards/combined_reward/mean": 7.937563562393189, "rewards/combined_reward/std": 2.3161841511726378, "step": 58140 }, { "clip_ratio/high_max": 0.0016996684030649097, "clip_ratio/high_mean": 0.0009795525606023149, "clip_ratio/low_mean": 0.0006472200422043291, "clip_ratio/low_min": 0.00018722696501451234, "clip_ratio/region_mean": 0.0016267725986836012, "completions/clipped_ratio": 0.0, "completions/max_length": 549.6, "completions/max_terminated_length": 549.6, "completions/mean_length": 216.96198577880858, "completions/mean_terminated_length": 216.96198577880858, "completions/min_length": 46.1, "completions/min_terminated_length": 46.1, "entropy": 0.025130295924221478, "epoch": 35.59993878175696, "frac_reward_zero_std": 0.13750000298023224, "grad_norm": 0.0260009765625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5955255795.0, "reward": 8.082403993606567, "reward_std": 0.3639713853597641, "rewards/combined_reward/mean": 8.082403707504273, "rewards/combined_reward/std": 2.1186797976493836, "step": 58170 }, { "clip_ratio/high_max": 0.0018055831615735466, "clip_ratio/high_mean": 0.0009776291009378232, "clip_ratio/low_mean": 0.0006729176800339095, "clip_ratio/low_min": 0.00014555926827597432, "clip_ratio/region_mean": 0.0016505467705428601, "completions/clipped_ratio": 0.0, "completions/max_length": 523.2, "completions/max_terminated_length": 523.2, "completions/mean_length": 199.58542175292968, "completions/mean_terminated_length": 199.58542175292968, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028147261093060174, "epoch": 35.61830425466789, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.004547119140625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 5958322363.0, "reward": 8.135408782958985, "reward_std": 0.2539185494184494, "rewards/combined_reward/mean": 8.135408544540406, "rewards/combined_reward/std": 2.1767664670944216, "step": 58200 }, { "clip_ratio/high_max": 0.002044060611418293, "clip_ratio/high_mean": 0.0011561972275861384, "clip_ratio/low_mean": 0.0008159063943215491, "clip_ratio/low_min": 0.00022642126132268458, "clip_ratio/region_mean": 0.001972103642280369, "completions/clipped_ratio": 0.0, "completions/max_length": 666.1, "completions/max_terminated_length": 666.1, "completions/mean_length": 214.53229675292968, "completions/mean_terminated_length": 214.53229675292968, "completions/min_length": 47.5, "completions/min_terminated_length": 47.5, "entropy": 0.028951438221459586, "epoch": 35.63666972757882, "frac_reward_zero_std": 0.12916666977107524, "grad_norm": 0.0308837890625, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 5961446327.0, "reward": 8.14673047065735, "reward_std": 0.3342940382659435, "rewards/combined_reward/mean": 8.146729946136475, "rewards/combined_reward/std": 2.165462625026703, "step": 58230 }, { "clip_ratio/high_max": 0.002058689875411801, "clip_ratio/high_mean": 0.001096929287329355, "clip_ratio/low_mean": 0.0007697250655231377, "clip_ratio/low_min": 0.00018778367593768052, "clip_ratio/region_mean": 0.0018666543750441634, "completions/clipped_ratio": 0.0, "completions/max_length": 566.7, "completions/max_terminated_length": 566.7, "completions/mean_length": 204.31641235351563, "completions/mean_terminated_length": 204.31641235351563, "completions/min_length": 37.5, "completions/min_terminated_length": 37.5, "entropy": 0.02795482734994342, "epoch": 35.65503520048975, "frac_reward_zero_std": 0.18750000447034837, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5964531062.0, "reward": 8.174510717391968, "reward_std": 0.27365431785583494, "rewards/combined_reward/mean": 8.174510669708251, "rewards/combined_reward/std": 2.0769260466098785, "step": 58260 }, { "clip_ratio/high_max": 0.002088212117087096, "clip_ratio/high_mean": 0.0012088867488879865, "clip_ratio/low_mean": 0.0008143997587467311, "clip_ratio/low_min": 0.0002310869550759283, "clip_ratio/region_mean": 0.0020232864995098983, "completions/clipped_ratio": 0.0, "completions/max_length": 557.1, "completions/max_terminated_length": 557.1, "completions/mean_length": 195.0679733276367, "completions/mean_terminated_length": 195.0679733276367, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.02977488610583047, "epoch": 35.67340067340067, "frac_reward_zero_std": 0.19583333991467952, "grad_norm": 0.0098876953125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5967580283.0, "reward": 7.827500247955323, "reward_std": 0.32800128012895585, "rewards/combined_reward/mean": 7.827500057220459, "rewards/combined_reward/std": 2.303332281112671, "step": 58290 }, { "clip_ratio/high_max": 0.001697096449546128, "clip_ratio/high_mean": 0.0009830239621805958, "clip_ratio/low_mean": 0.0006474485907043951, "clip_ratio/low_min": 0.0001826018303593931, "clip_ratio/region_mean": 0.0016304725422135865, "completions/clipped_ratio": 0.0, "completions/max_length": 563.1, "completions/max_terminated_length": 563.1, "completions/mean_length": 201.2304718017578, "completions/mean_terminated_length": 201.2304718017578, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.025334436322251956, "epoch": 35.6917661463116, "frac_reward_zero_std": 0.20416667088866233, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5970653168.0, "reward": 8.137949562072754, "reward_std": 0.25143377780914306, "rewards/combined_reward/mean": 8.137949323654174, "rewards/combined_reward/std": 2.1986531257629394, "step": 58320 }, { "clip_ratio/high_max": 0.0021425577127956786, "clip_ratio/high_mean": 0.001118286301061744, "clip_ratio/low_mean": 0.0007803401970401561, "clip_ratio/low_min": 0.00017014008214270385, "clip_ratio/region_mean": 0.0018986264942213893, "completions/clipped_ratio": 0.0, "completions/max_length": 504.7, "completions/max_terminated_length": 504.7, "completions/mean_length": 206.98802642822267, "completions/mean_terminated_length": 206.98802642822267, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.028743829927407205, "epoch": 35.71013161922253, "frac_reward_zero_std": 0.10416666828095913, "grad_norm": 0.0076904296875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 5973748162.0, "reward": 8.322397994995118, "reward_std": 0.3042739987373352, "rewards/combined_reward/mean": 8.32239751815796, "rewards/combined_reward/std": 1.868950641155243, "step": 58350 }, { "clip_ratio/high_max": 0.0019142063208467638, "clip_ratio/high_mean": 0.0011203762664081297, "clip_ratio/low_mean": 0.0007644962700093553, "clip_ratio/low_min": 0.00019044067982273798, "clip_ratio/region_mean": 0.001884872541995719, "completions/clipped_ratio": 0.0, "completions/max_length": 568.8, "completions/max_terminated_length": 568.8, "completions/mean_length": 214.63255767822267, "completions/mean_terminated_length": 214.63255767822267, "completions/min_length": 51.1, "completions/min_terminated_length": 51.1, "entropy": 0.02875780618439118, "epoch": 35.72849709213346, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.019775390625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5976872511.0, "reward": 8.070811700820922, "reward_std": 0.36769499629735947, "rewards/combined_reward/mean": 8.070811462402343, "rewards/combined_reward/std": 2.3453264951705934, "step": 58380 }, { "clip_ratio/high_max": 0.001803799108893145, "clip_ratio/high_mean": 0.0009706246055429801, "clip_ratio/low_mean": 0.0007436419948741483, "clip_ratio/low_min": 0.0002157474703077848, "clip_ratio/region_mean": 0.0017142665892606601, "completions/clipped_ratio": 0.0, "completions/max_length": 592.2, "completions/max_terminated_length": 592.2, "completions/mean_length": 206.5901077270508, "completions/mean_terminated_length": 206.5901077270508, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.02567890935267011, "epoch": 35.74686256504438, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.005462646484375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 5979965977.0, "reward": 8.145005321502685, "reward_std": 0.30244831591844556, "rewards/combined_reward/mean": 8.14500527381897, "rewards/combined_reward/std": 2.243097162246704, "step": 58410 }, { "clip_ratio/high_max": 0.0021893536584684627, "clip_ratio/high_mean": 0.0012813348133931869, "clip_ratio/low_mean": 0.000759318553173216, "clip_ratio/low_min": 0.0002577385758437837, "clip_ratio/region_mean": 0.002040653393487446, "completions/clipped_ratio": 0.0, "completions/max_length": 547.8, "completions/max_terminated_length": 547.8, "completions/mean_length": 211.0510467529297, "completions/mean_terminated_length": 211.0510467529297, "completions/min_length": 47.2, "completions/min_terminated_length": 47.2, "entropy": 0.0302005749506255, "epoch": 35.76522803795531, "frac_reward_zero_std": 0.10833333507180214, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5983076573.0, "reward": 8.097003030776978, "reward_std": 0.3264589883387089, "rewards/combined_reward/mean": 8.09700255393982, "rewards/combined_reward/std": 2.226944053173065, "step": 58440 }, { "clip_ratio/high_max": 0.0018631299540478115, "clip_ratio/high_mean": 0.0010601913743206145, "clip_ratio/low_mean": 0.000803259326130501, "clip_ratio/low_min": 0.0002649839368435399, "clip_ratio/region_mean": 0.0018634506976619983, "completions/clipped_ratio": 0.0, "completions/max_length": 541.4, "completions/max_terminated_length": 541.4, "completions/mean_length": 205.9992248535156, "completions/mean_terminated_length": 205.9992248535156, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.027481937035918235, "epoch": 35.78359351086624, "frac_reward_zero_std": 0.1750000037252903, "grad_norm": 0.0277099609375, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 5986167770.0, "reward": 8.24259386062622, "reward_std": 0.22688871920108794, "rewards/combined_reward/mean": 8.242593431472779, "rewards/combined_reward/std": 2.0948635220527647, "step": 58470 }, { "clip_ratio/high_max": 0.0019287260736746247, "clip_ratio/high_mean": 0.0010704871764270743, "clip_ratio/low_mean": 0.0008264510994195007, "clip_ratio/low_min": 0.00024808772795950065, "clip_ratio/region_mean": 0.0018969382717235325, "completions/clipped_ratio": 0.0, "completions/max_length": 571.2, "completions/max_terminated_length": 571.2, "completions/mean_length": 204.63177337646485, "completions/mean_terminated_length": 204.63177337646485, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.028829883923754095, "epoch": 35.80195898377717, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 5989253716.0, "reward": 8.076279592514037, "reward_std": 0.276307737827301, "rewards/combined_reward/mean": 8.076279401779175, "rewards/combined_reward/std": 2.1635976433753967, "step": 58500 }, { "clip_ratio/high_max": 0.0020998068546759898, "clip_ratio/high_mean": 0.0011752354432246648, "clip_ratio/low_mean": 0.0008150382188129394, "clip_ratio/low_min": 0.00031099138917246215, "clip_ratio/region_mean": 0.0019902736671307745, "completions/clipped_ratio": 0.0, "completions/max_length": 594.1, "completions/max_terminated_length": 594.1, "completions/mean_length": 201.44427642822265, "completions/mean_terminated_length": 201.44427642822265, "completions/min_length": 40.2, "completions/min_terminated_length": 40.2, "entropy": 0.02852385598234832, "epoch": 35.82032445668809, "frac_reward_zero_std": 0.22083334177732467, "grad_norm": 0.023193359375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 5992327422.0, "reward": 7.838583469390869, "reward_std": 0.307876855134964, "rewards/combined_reward/mean": 7.838583087921142, "rewards/combined_reward/std": 2.5936230421066284, "step": 58530 }, { "clip_ratio/high_max": 0.0025970980311588693, "clip_ratio/high_mean": 0.001378025640588021, "clip_ratio/low_mean": 0.0008066363376807809, "clip_ratio/low_min": 0.00018593595935575044, "clip_ratio/region_mean": 0.0021846619638381526, "completions/clipped_ratio": 0.0, "completions/max_length": 604.4, "completions/max_terminated_length": 604.4, "completions/mean_length": 200.91068267822266, "completions/mean_terminated_length": 200.91068267822266, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.032743883427853386, "epoch": 35.83868992959902, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.016845703125, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 5995399079.0, "reward": 8.273192024230957, "reward_std": 0.36290671229362487, "rewards/combined_reward/mean": 8.273191785812378, "rewards/combined_reward/std": 2.0691007494926454, "step": 58560 }, { "clip_ratio/high_max": 0.001455416052097765, "clip_ratio/high_mean": 0.0008651495497664049, "clip_ratio/low_mean": 0.000791668752329618, "clip_ratio/low_min": 0.00022603718025493436, "clip_ratio/region_mean": 0.001656818295790193, "completions/clipped_ratio": 0.0, "completions/max_length": 572.4, "completions/max_terminated_length": 572.4, "completions/mean_length": 202.18021545410156, "completions/mean_terminated_length": 202.18021545410156, "completions/min_length": 42.9, "completions/min_terminated_length": 42.9, "entropy": 0.027058313429976503, "epoch": 35.85705540250995, "frac_reward_zero_std": 0.20833333879709243, "grad_norm": 0.00555419921875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 5998475611.0, "reward": 8.350877523422241, "reward_std": 0.2393802970647812, "rewards/combined_reward/mean": 8.350877332687379, "rewards/combined_reward/std": 1.8630862712860108, "step": 58590 }, { "clip_ratio/high_max": 0.0018869956494503035, "clip_ratio/high_mean": 0.0011990446885950707, "clip_ratio/low_mean": 0.0007595546308342212, "clip_ratio/low_min": 0.00023067512302077375, "clip_ratio/region_mean": 0.0019585993368915902, "completions/clipped_ratio": 0.0, "completions/max_length": 614.8, "completions/max_terminated_length": 614.8, "completions/mean_length": 211.01562957763673, "completions/mean_terminated_length": 211.01562957763673, "completions/min_length": 47.3, "completions/min_terminated_length": 47.3, "entropy": 0.029882529323610168, "epoch": 35.87542087542087, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.00799560546875, "learning_rate": 5e-06, "loss": 0.0022, "num_tokens": 6001586071.0, "reward": 8.228357601165772, "reward_std": 0.3520701751112938, "rewards/combined_reward/mean": 8.228357410430908, "rewards/combined_reward/std": 2.1284058928489684, "step": 58620 }, { "clip_ratio/high_max": 0.00169866013796612, "clip_ratio/high_mean": 0.0009917669456626754, "clip_ratio/low_mean": 0.0007548889581206216, "clip_ratio/low_min": 0.00027128185150407564, "clip_ratio/region_mean": 0.0017466559089371002, "completions/clipped_ratio": 0.0, "completions/max_length": 596.4, "completions/max_terminated_length": 596.4, "completions/mean_length": 210.98750762939454, "completions/mean_terminated_length": 210.98750762939454, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "entropy": 0.0268357133337607, "epoch": 35.8937863483318, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 6004696423.0, "reward": 8.167819547653199, "reward_std": 0.303341318666935, "rewards/combined_reward/mean": 8.16781930923462, "rewards/combined_reward/std": 2.097920799255371, "step": 58650 }, { "clip_ratio/high_max": 0.0023130890496152762, "clip_ratio/high_mean": 0.001259290491240487, "clip_ratio/low_mean": 0.000763113564850452, "clip_ratio/low_min": 0.00020498744003513518, "clip_ratio/region_mean": 0.0020224040519678963, "completions/clipped_ratio": 0.0, "completions/max_length": 620.4, "completions/max_terminated_length": 620.4, "completions/mean_length": 197.05442962646484, "completions/mean_terminated_length": 197.05442962646484, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.03021786993679901, "epoch": 35.91215182124273, "frac_reward_zero_std": 0.18333333879709243, "grad_norm": 0.00872802734375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 6007753272.0, "reward": 8.053998661041259, "reward_std": 0.3250822752714157, "rewards/combined_reward/mean": 8.053998613357544, "rewards/combined_reward/std": 2.019763100147247, "step": 58680 }, { "clip_ratio/high_max": 0.001879277022089809, "clip_ratio/high_mean": 0.0010927350806014147, "clip_ratio/low_mean": 0.0007805600164526064, "clip_ratio/low_min": 0.00022534362360602245, "clip_ratio/region_mean": 0.0018732951141525215, "completions/clipped_ratio": 0.0, "completions/max_length": 686.4, "completions/max_terminated_length": 686.4, "completions/mean_length": 205.63386077880858, "completions/mean_terminated_length": 205.63386077880858, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.031014345043028396, "epoch": 35.93051729415366, "frac_reward_zero_std": 0.19583333730697633, "grad_norm": 0.0201416015625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 6010843066.0, "reward": 8.25898118019104, "reward_std": 0.2952753648161888, "rewards/combined_reward/mean": 8.258980798721314, "rewards/combined_reward/std": 2.103926867246628, "step": 58710 }, { "clip_ratio/high_max": 0.0019136805773693293, "clip_ratio/high_mean": 0.001134825183544308, "clip_ratio/low_mean": 0.0008377941644236368, "clip_ratio/low_min": 0.00025676241299758357, "clip_ratio/region_mean": 0.001972619347119083, "completions/clipped_ratio": 0.0, "completions/max_length": 560.3, "completions/max_terminated_length": 560.3, "completions/mean_length": 201.63489990234376, "completions/mean_terminated_length": 201.63489990234376, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "entropy": 0.02653192662789176, "epoch": 35.94888276706458, "frac_reward_zero_std": 0.20833333693444728, "grad_norm": 0.01513671875, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 6013917504.0, "reward": 7.863433027267456, "reward_std": 0.2795752339065075, "rewards/combined_reward/mean": 7.863432741165161, "rewards/combined_reward/std": 2.4298841834068297, "step": 58740 }, { "clip_ratio/high_max": 0.0018866606570857888, "clip_ratio/high_mean": 0.0011777122357064702, "clip_ratio/low_mean": 0.0006287622500773675, "clip_ratio/low_min": 0.0001078975741014195, "clip_ratio/region_mean": 0.00180647448481371, "completions/clipped_ratio": 0.0, "completions/max_length": 618.4, "completions/max_terminated_length": 618.4, "completions/mean_length": 201.80286865234376, "completions/mean_terminated_length": 201.80286865234376, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.027525928538913527, "epoch": 35.96724823997551, "frac_reward_zero_std": 0.20000000521540642, "grad_norm": 0.0130615234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 6016992587.0, "reward": 7.941787528991699, "reward_std": 0.28545535951852796, "rewards/combined_reward/mean": 7.941787433624268, "rewards/combined_reward/std": 2.517294633388519, "step": 58770 }, { "clip_ratio/high_max": 0.0015462223692641905, "clip_ratio/high_mean": 0.0009230589186093615, "clip_ratio/low_mean": 0.0007213171496308254, "clip_ratio/low_min": 0.000253868120489642, "clip_ratio/region_mean": 0.0016443760638746123, "completions/clipped_ratio": 0.0, "completions/max_length": 565.6, "completions/max_terminated_length": 565.6, "completions/mean_length": 204.16588897705077, "completions/mean_terminated_length": 204.16588897705077, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.02545602071719865, "epoch": 35.98561371288644, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.0189208984375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6020076744.0, "reward": 8.186634922027588, "reward_std": 0.28799650371074675, "rewards/combined_reward/mean": 8.186634588241578, "rewards/combined_reward/std": 2.0117213249206545, "step": 58800 }, { "clip_ratio/high_max": 0.0019899495031883117, "clip_ratio/high_mean": 0.001108130150747516, "clip_ratio/low_mean": 0.0008887549054515014, "clip_ratio/low_min": 0.00027016047235639367, "clip_ratio/region_mean": 0.0019968850452254418, "completions/clipped_ratio": 0.0, "completions/max_length": 608.4, "completions/max_terminated_length": 608.4, "completions/mean_length": 206.97396697998047, "completions/mean_terminated_length": 206.97396697998047, "completions/min_length": 48.8, "completions/min_terminated_length": 48.8, "entropy": 0.027455673308768233, "epoch": 36.00428527701255, "frac_reward_zero_std": 0.15416667014360427, "grad_norm": 0.0242919921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6023171684.0, "reward": 8.29338517189026, "reward_std": 0.28080509454011915, "rewards/combined_reward/mean": 8.293385076522828, "rewards/combined_reward/std": 2.154559552669525, "step": 58830 }, { "clip_ratio/high_max": 0.00227828588540433, "clip_ratio/high_mean": 0.0013536750586354174, "clip_ratio/low_mean": 0.0011755667206064875, "clip_ratio/low_min": 0.00045215084489124514, "clip_ratio/region_mean": 0.0025292417863965965, "completions/clipped_ratio": 0.0, "completions/max_length": 523.6, "completions/max_terminated_length": 523.6, "completions/mean_length": 193.5955780029297, "completions/mean_terminated_length": 193.5955780029297, "completions/min_length": 39.6, "completions/min_terminated_length": 39.6, "entropy": 0.029765001932779946, "epoch": 36.02265074992348, "frac_reward_zero_std": 0.17500000298023224, "grad_norm": 0.0155029296875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 6026215251.0, "reward": 7.973861217498779, "reward_std": 0.3231509640812874, "rewards/combined_reward/mean": 7.9738609790802, "rewards/combined_reward/std": 2.3017678260803223, "step": 58860 }, { "clip_ratio/high_max": 0.0028846031690288026, "clip_ratio/high_mean": 0.0017476357833705455, "clip_ratio/low_mean": 0.0010196733025319797, "clip_ratio/low_min": 0.00032416780741186815, "clip_ratio/region_mean": 0.0027673091162190152, "completions/clipped_ratio": 0.0, "completions/max_length": 535.8, "completions/max_terminated_length": 535.8, "completions/mean_length": 213.28411865234375, "completions/mean_terminated_length": 213.28411865234375, "completions/min_length": 37.3, "completions/min_terminated_length": 37.3, "entropy": 0.03277971895877272, "epoch": 36.04101622283441, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0238037109375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 6029334422.0, "reward": 8.099203443527221, "reward_std": 0.3201058655977249, "rewards/combined_reward/mean": 8.099203157424927, "rewards/combined_reward/std": 2.205899250507355, "step": 58890 }, { "clip_ratio/high_max": 0.0018361923129608234, "clip_ratio/high_mean": 0.0010569642384022397, "clip_ratio/low_mean": 0.000924961396594881, "clip_ratio/low_min": 0.00019505939077741157, "clip_ratio/region_mean": 0.001981925646396121, "completions/clipped_ratio": 0.0, "completions/max_length": 574.5, "completions/max_terminated_length": 574.5, "completions/mean_length": 199.57760925292968, "completions/mean_terminated_length": 199.57760925292968, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.025038917611042658, "epoch": 36.05938169574533, "frac_reward_zero_std": 0.12916667051613331, "grad_norm": 0.0211181640625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 6032400960.0, "reward": 8.174172353744506, "reward_std": 0.2663387030363083, "rewards/combined_reward/mean": 8.174172067642212, "rewards/combined_reward/std": 2.0687795758247374, "step": 58920 }, { "clip_ratio/high_max": 0.0019833631294507844, "clip_ratio/high_mean": 0.0011401097630975225, "clip_ratio/low_mean": 0.0009045881419297075, "clip_ratio/low_min": 0.0002388874347767948, "clip_ratio/region_mean": 0.0020446978870798677, "completions/clipped_ratio": 0.0, "completions/max_length": 553.5, "completions/max_terminated_length": 553.5, "completions/mean_length": 197.2593795776367, "completions/mean_terminated_length": 197.2593795776367, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.027235618342335025, "epoch": 36.07774716865626, "frac_reward_zero_std": 0.11250000149011612, "grad_norm": 0.004058837890625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6035458596.0, "reward": 8.25793523788452, "reward_std": 0.3360472142696381, "rewards/combined_reward/mean": 8.257935094833375, "rewards/combined_reward/std": 2.096966230869293, "step": 58950 }, { "clip_ratio/high_max": 0.0025496012439058784, "clip_ratio/high_mean": 0.0014460502217843896, "clip_ratio/low_mean": 0.0009936435338507483, "clip_ratio/low_min": 0.0003300787393527571, "clip_ratio/region_mean": 0.002439693734534861, "completions/clipped_ratio": 0.0, "completions/max_length": 619.1, "completions/max_terminated_length": 619.1, "completions/mean_length": 208.84193115234376, "completions/mean_terminated_length": 208.84193115234376, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.029284641441578666, "epoch": 36.09611264156719, "frac_reward_zero_std": 0.15000000707805156, "grad_norm": 0.0166015625, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 6038560709.0, "reward": 8.0778724193573, "reward_std": 0.32722203731536864, "rewards/combined_reward/mean": 8.077872228622436, "rewards/combined_reward/std": 2.0856422543525697, "step": 58980 }, { "clip_ratio/high_max": 0.002265092897384117, "clip_ratio/high_mean": 0.0013172206438563686, "clip_ratio/low_mean": 0.001038606231547116, "clip_ratio/low_min": 0.0003655330806698961, "clip_ratio/region_mean": 0.002355826892987049, "completions/clipped_ratio": 0.0, "completions/max_length": 741.6, "completions/max_terminated_length": 741.6, "completions/mean_length": 213.83490142822265, "completions/mean_terminated_length": 213.83490142822265, "completions/min_length": 46.3, "completions/min_terminated_length": 46.3, "entropy": 0.029618349283312757, "epoch": 36.11447811447812, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6041681995.0, "reward": 7.8732506275177006, "reward_std": 0.31377298980951307, "rewards/combined_reward/mean": 7.8732503890991214, "rewards/combined_reward/std": 2.4610227704048158, "step": 59010 }, { "clip_ratio/high_max": 0.002888957833541402, "clip_ratio/high_mean": 0.0017191678084297261, "clip_ratio/low_mean": 0.0010253966469463195, "clip_ratio/low_min": 0.0002499035889438043, "clip_ratio/region_mean": 0.002744564426636013, "completions/clipped_ratio": 0.0, "completions/max_length": 593.4, "completions/max_terminated_length": 593.4, "completions/mean_length": 196.0197952270508, "completions/mean_terminated_length": 196.0197952270508, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.03360183920400838, "epoch": 36.13284358738904, "frac_reward_zero_std": 0.20416667610406875, "grad_norm": 0.00909423828125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6044734871.0, "reward": 8.05411524772644, "reward_std": 0.26743529438972474, "rewards/combined_reward/mean": 8.054114770889282, "rewards/combined_reward/std": 2.3213815331459045, "step": 59040 }, { "clip_ratio/high_max": 0.0028384887215603762, "clip_ratio/high_mean": 0.0014563399473748481, "clip_ratio/low_mean": 0.0009422555786538094, "clip_ratio/low_min": 0.0002745083901875963, "clip_ratio/region_mean": 0.0023985955068686355, "completions/clipped_ratio": 0.0, "completions/max_length": 585.9, "completions/max_terminated_length": 585.9, "completions/mean_length": 198.4205764770508, "completions/mean_terminated_length": 198.4205764770508, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.027805205434560776, "epoch": 36.15120906029997, "frac_reward_zero_std": 0.1541666705161333, "grad_norm": 0.01531982421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6047796966.0, "reward": 8.006792306900024, "reward_std": 0.3269555173814297, "rewards/combined_reward/mean": 8.006792211532593, "rewards/combined_reward/std": 2.2308155417442324, "step": 59070 }, { "clip_ratio/high_max": 0.0023031696512286243, "clip_ratio/high_mean": 0.0012656456528929994, "clip_ratio/low_mean": 0.0010134363573646018, "clip_ratio/low_min": 0.00027319916771375576, "clip_ratio/region_mean": 0.002279082019716346, "completions/clipped_ratio": 0.0, "completions/max_length": 607.9, "completions/max_terminated_length": 607.9, "completions/mean_length": 208.76849365234375, "completions/mean_terminated_length": 208.76849365234375, "completions/min_length": 42.7, "completions/min_terminated_length": 42.7, "entropy": 0.02787897295008103, "epoch": 36.1695745332109, "frac_reward_zero_std": 0.13750000335276127, "grad_norm": 0.0096435546875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6050898797.0, "reward": 8.064391040802002, "reward_std": 0.28501408696174624, "rewards/combined_reward/mean": 8.064390754699707, "rewards/combined_reward/std": 2.2801796197891235, "step": 59100 }, { "clip_ratio/high_max": 0.0020932804104328777, "clip_ratio/high_mean": 0.0012603404313267675, "clip_ratio/low_mean": 0.000745501216078992, "clip_ratio/low_min": 0.00013541468118395035, "clip_ratio/region_mean": 0.002005841654802983, "completions/clipped_ratio": 0.0, "completions/max_length": 510.1, "completions/max_terminated_length": 510.1, "completions/mean_length": 200.88516082763672, "completions/mean_terminated_length": 200.88516082763672, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.025851614214479922, "epoch": 36.18794000612183, "frac_reward_zero_std": 0.17916667014360427, "grad_norm": 0.01153564453125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6053970356.0, "reward": 8.11077160835266, "reward_std": 0.28851599544286727, "rewards/combined_reward/mean": 8.110771322250367, "rewards/combined_reward/std": 2.2486500263214113, "step": 59130 }, { "clip_ratio/high_max": 0.0021131254160233464, "clip_ratio/high_mean": 0.001205100195875275, "clip_ratio/low_mean": 0.0009316409182550463, "clip_ratio/low_min": 0.00023974263167474418, "clip_ratio/region_mean": 0.002136741109522215, "completions/clipped_ratio": 0.0, "completions/max_length": 578.8, "completions/max_terminated_length": 578.8, "completions/mean_length": 194.4940185546875, "completions/mean_terminated_length": 194.4940185546875, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.026668481575325132, "epoch": 36.20630547903275, "frac_reward_zero_std": 0.1958333384245634, "grad_norm": 0.00604248046875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 6057017373.0, "reward": 8.094778156280517, "reward_std": 0.2293913781642914, "rewards/combined_reward/mean": 8.094777917861938, "rewards/combined_reward/std": 2.160959577560425, "step": 59160 }, { "clip_ratio/high_max": 0.002450356030991922, "clip_ratio/high_mean": 0.0014303756859590066, "clip_ratio/low_mean": 0.0009227481051008606, "clip_ratio/low_min": 0.00027107526305674887, "clip_ratio/region_mean": 0.0023531237966381013, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 197.89688110351562, "completions/mean_terminated_length": 197.89688110351562, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.031018313222254314, "epoch": 36.22467095194368, "frac_reward_zero_std": 0.17916667088866234, "grad_norm": 0.0208740234375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6060077457.0, "reward": 8.019950532913208, "reward_std": 0.3142254650592804, "rewards/combined_reward/mean": 8.01995029449463, "rewards/combined_reward/std": 2.4853888511657716, "step": 59190 }, { "clip_ratio/high_max": 0.002551426463954461, "clip_ratio/high_mean": 0.0014346192794619129, "clip_ratio/low_mean": 0.0009649345904108487, "clip_ratio/low_min": 0.0002644072817929555, "clip_ratio/region_mean": 0.002399553868842001, "completions/clipped_ratio": 0.0, "completions/max_length": 582.9, "completions/max_terminated_length": 582.9, "completions/mean_length": 197.87318267822266, "completions/mean_terminated_length": 197.87318267822266, "completions/min_length": 38.8, "completions/min_terminated_length": 38.8, "entropy": 0.029009573673829437, "epoch": 36.24303642485461, "frac_reward_zero_std": 0.1583333369344473, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6063137450.0, "reward": 8.174415826797485, "reward_std": 0.2805274873971939, "rewards/combined_reward/mean": 8.17441554069519, "rewards/combined_reward/std": 2.049212908744812, "step": 59220 }, { "clip_ratio/high_max": 0.0022345244394576485, "clip_ratio/high_mean": 0.0012638062908081338, "clip_ratio/low_mean": 0.000859876157604352, "clip_ratio/low_min": 0.0001996376035094727, "clip_ratio/region_mean": 0.002123682451201603, "completions/clipped_ratio": 0.0, "completions/max_length": 636.9, "completions/max_terminated_length": 636.9, "completions/mean_length": 202.8190139770508, "completions/mean_terminated_length": 202.8190139770508, "completions/min_length": 44.6, "completions/min_terminated_length": 44.6, "entropy": 0.026997826357061664, "epoch": 36.26140189776554, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0087890625, "learning_rate": 5e-06, "loss": 0.0005, "num_tokens": 6066216435.0, "reward": 8.123691654205322, "reward_std": 0.25792121514678, "rewards/combined_reward/mean": 8.123691511154174, "rewards/combined_reward/std": 2.1702364802360536, "step": 59250 }, { "clip_ratio/high_max": 0.00238043433103788, "clip_ratio/high_mean": 0.0013224431919904116, "clip_ratio/low_mean": 0.0010904172513013085, "clip_ratio/low_min": 0.0003482558977945397, "clip_ratio/region_mean": 0.002412860464877061, "completions/clipped_ratio": 0.0, "completions/max_length": 622.3, "completions/max_terminated_length": 622.3, "completions/mean_length": 191.49349365234374, "completions/mean_terminated_length": 191.49349365234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.031764597973475854, "epoch": 36.27976737067646, "frac_reward_zero_std": 0.20833333656191827, "grad_norm": 0.01422119140625, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 6069251930.0, "reward": 8.102919149398804, "reward_std": 0.21155069544911384, "rewards/combined_reward/mean": 8.102919101715088, "rewards/combined_reward/std": 2.232466697692871, "step": 59280 }, { "clip_ratio/high_max": 0.002520994173149423, "clip_ratio/high_mean": 0.001376704147211664, "clip_ratio/low_mean": 0.0008339155184633758, "clip_ratio/low_min": 0.000269884176668711, "clip_ratio/region_mean": 0.002210619650092364, "completions/clipped_ratio": 0.0, "completions/max_length": 635.9, "completions/max_terminated_length": 635.9, "completions/mean_length": 202.52344512939453, "completions/mean_terminated_length": 202.52344512939453, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.027398382360115648, "epoch": 36.29813284358739, "frac_reward_zero_std": 0.1833333417773247, "grad_norm": 0.036376953125, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6072329780.0, "reward": 7.887678098678589, "reward_std": 0.35326572209596635, "rewards/combined_reward/mean": 7.887677812576294, "rewards/combined_reward/std": 2.3573455214500427, "step": 59310 }, { "clip_ratio/high_max": 0.002385263481604246, "clip_ratio/high_mean": 0.00141378562572451, "clip_ratio/low_mean": 0.0008745474477715713, "clip_ratio/low_min": 0.00023846316156171573, "clip_ratio/region_mean": 0.0022883330520320063, "completions/clipped_ratio": 0.0, "completions/max_length": 575.8, "completions/max_terminated_length": 575.8, "completions/mean_length": 199.0065139770508, "completions/mean_terminated_length": 199.0065139770508, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.027843244792893528, "epoch": 36.31649831649832, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.028076171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6075394125.0, "reward": 8.298830270767212, "reward_std": 0.394063712656498, "rewards/combined_reward/mean": 8.29883017539978, "rewards/combined_reward/std": 2.0388797879219056, "step": 59340 }, { "clip_ratio/high_max": 0.0023319516697180615, "clip_ratio/high_mean": 0.0013749408063934729, "clip_ratio/low_mean": 0.0009976472151417208, "clip_ratio/low_min": 0.0002863201589207165, "clip_ratio/region_mean": 0.002372588004800491, "completions/clipped_ratio": 0.0, "completions/max_length": 672.5, "completions/max_terminated_length": 672.5, "completions/mean_length": 206.3869842529297, "completions/mean_terminated_length": 206.3869842529297, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.03057534354738891, "epoch": 36.33486378940924, "frac_reward_zero_std": 0.15416667200624942, "grad_norm": 0.0157470703125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6078486811.0, "reward": 8.127082014083863, "reward_std": 0.3293479114770889, "rewards/combined_reward/mean": 8.127081537246704, "rewards/combined_reward/std": 2.2729412794113157, "step": 59370 }, { "clip_ratio/high_max": 0.0023729219848367695, "clip_ratio/high_mean": 0.0012822565389797092, "clip_ratio/low_mean": 0.000785266951546267, "clip_ratio/low_min": 0.0002176814574340824, "clip_ratio/region_mean": 0.0020675234942852208, "completions/clipped_ratio": 0.0, "completions/max_length": 602.1, "completions/max_terminated_length": 602.1, "completions/mean_length": 202.51380767822266, "completions/mean_terminated_length": 202.51380767822266, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.029217217536643147, "epoch": 36.35322926232017, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.00994873046875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6081564624.0, "reward": 8.002309131622315, "reward_std": 0.31595503240823747, "rewards/combined_reward/mean": 8.002308893203736, "rewards/combined_reward/std": 2.384176325798035, "step": 59400 }, { "clip_ratio/high_max": 0.0025424435203603934, "clip_ratio/high_mean": 0.0015138392952697662, "clip_ratio/low_mean": 0.0010223757784842747, "clip_ratio/low_min": 0.0002805850187238927, "clip_ratio/region_mean": 0.002536215091337605, "completions/clipped_ratio": 0.0, "completions/max_length": 616.6, "completions/max_terminated_length": 616.6, "completions/mean_length": 201.15365142822264, "completions/mean_terminated_length": 201.15365142822264, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.03000254799456646, "epoch": 36.3715947352311, "frac_reward_zero_std": 0.16250000447034835, "grad_norm": 0.01519775390625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6084637214.0, "reward": 8.17244815826416, "reward_std": 0.27762170881032944, "rewards/combined_reward/mean": 8.172448015213012, "rewards/combined_reward/std": 2.1517303347587586, "step": 59430 }, { "clip_ratio/high_max": 0.0024097755134183294, "clip_ratio/high_mean": 0.0014000089899733818, "clip_ratio/low_mean": 0.0011474638204163057, "clip_ratio/low_min": 0.0003224311550487376, "clip_ratio/region_mean": 0.0025474728230619804, "completions/clipped_ratio": 0.0, "completions/max_length": 567.5, "completions/max_terminated_length": 567.5, "completions/mean_length": 205.98203735351564, "completions/mean_terminated_length": 205.98203735351564, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.030891321770225962, "epoch": 36.38996020814203, "frac_reward_zero_std": 0.1708333369344473, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6087728345.0, "reward": 7.833529806137085, "reward_std": 0.2874318078160286, "rewards/combined_reward/mean": 7.833529710769653, "rewards/combined_reward/std": 2.4057593941688538, "step": 59460 }, { "clip_ratio/high_max": 0.002158004772597148, "clip_ratio/high_mean": 0.001176839887072371, "clip_ratio/low_mean": 0.0009717492778387774, "clip_ratio/low_min": 0.0002055676086456515, "clip_ratio/region_mean": 0.0021485891712169783, "completions/clipped_ratio": 0.0, "completions/max_length": 556.6, "completions/max_terminated_length": 556.6, "completions/mean_length": 206.53099670410157, "completions/mean_terminated_length": 206.53099670410157, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.02740613599307835, "epoch": 36.40832568105295, "frac_reward_zero_std": 0.1500000037252903, "grad_norm": 0.031494140625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6090821584.0, "reward": 8.045373344421387, "reward_std": 0.34455346316099167, "rewards/combined_reward/mean": 8.045373058319091, "rewards/combined_reward/std": 2.1637394547462465, "step": 59490 }, { "clip_ratio/high_max": 0.0022792028796781477, "clip_ratio/high_mean": 0.0012692424046690577, "clip_ratio/low_mean": 0.0008472910268513563, "clip_ratio/low_min": 0.00018546417195466346, "clip_ratio/region_mean": 0.0021165334449809356, "completions/clipped_ratio": 0.0, "completions/max_length": 561.5, "completions/max_terminated_length": 561.5, "completions/mean_length": 202.43437957763672, "completions/mean_terminated_length": 202.43437957763672, "completions/min_length": 44.3, "completions/min_terminated_length": 44.3, "entropy": 0.02506261969295641, "epoch": 36.42669115396388, "frac_reward_zero_std": 0.17083333767950534, "grad_norm": 0.012451171875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6093899092.0, "reward": 7.7610632419586185, "reward_std": 0.32210175395011903, "rewards/combined_reward/mean": 7.761062860488892, "rewards/combined_reward/std": 2.6435629606246946, "step": 59520 }, { "clip_ratio/high_max": 0.002113171110492355, "clip_ratio/high_mean": 0.0012333367194514722, "clip_ratio/low_mean": 0.0011107946794557697, "clip_ratio/low_min": 0.000357041455572471, "clip_ratio/region_mean": 0.0023441313877507735, "completions/clipped_ratio": 0.0, "completions/max_length": 594.3, "completions/max_terminated_length": 594.3, "completions/mean_length": 203.32474365234376, "completions/mean_terminated_length": 203.32474365234376, "completions/min_length": 49.1, "completions/min_terminated_length": 49.1, "entropy": 0.029935637069866062, "epoch": 36.44505662687481, "frac_reward_zero_std": 0.1416666701436043, "grad_norm": 0.0142822265625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6096980019.0, "reward": 8.119131565093994, "reward_std": 0.2768751889467239, "rewards/combined_reward/mean": 8.119131565093994, "rewards/combined_reward/std": 2.445688009262085, "step": 59550 }, { "clip_ratio/high_max": 0.0026801169386696227, "clip_ratio/high_mean": 0.0014667720410216135, "clip_ratio/low_mean": 0.0010938545442816879, "clip_ratio/low_min": 0.000290512550418498, "clip_ratio/region_mean": 0.002560626555835673, "completions/clipped_ratio": 0.0, "completions/max_length": 570.7, "completions/max_terminated_length": 570.7, "completions/mean_length": 199.77812957763672, "completions/mean_terminated_length": 199.77812957763672, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.030719522650664053, "epoch": 36.46342209978574, "frac_reward_zero_std": 0.2250000059604645, "grad_norm": 0.033203125, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 6100047327.0, "reward": 8.010748815536498, "reward_std": 0.4198643520474434, "rewards/combined_reward/mean": 8.010748624801636, "rewards/combined_reward/std": 2.276252293586731, "step": 59580 }, { "clip_ratio/high_max": 0.0027050037053413687, "clip_ratio/high_mean": 0.0015372980297494602, "clip_ratio/low_mean": 0.0010014970825674633, "clip_ratio/low_min": 0.00030085799468603603, "clip_ratio/region_mean": 0.0025387951209268067, "completions/clipped_ratio": 0.0, "completions/max_length": 616.9, "completions/max_terminated_length": 616.9, "completions/mean_length": 201.3585983276367, "completions/mean_terminated_length": 201.3585983276367, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.03160485712190469, "epoch": 36.48178757269666, "frac_reward_zero_std": 0.17916666865348815, "grad_norm": 0.0220947265625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6103120704.0, "reward": 7.999877452850342, "reward_std": 0.3501149833202362, "rewards/combined_reward/mean": 7.999877119064331, "rewards/combined_reward/std": 2.289943754673004, "step": 59610 }, { "clip_ratio/high_max": 0.002529298288815577, "clip_ratio/high_mean": 0.0014921460411035999, "clip_ratio/low_mean": 0.0009732709816792825, "clip_ratio/low_min": 0.000279854111916696, "clip_ratio/region_mean": 0.0024654170149005948, "completions/clipped_ratio": 0.0, "completions/max_length": 699.2, "completions/max_terminated_length": 699.2, "completions/mean_length": 206.55807647705078, "completions/mean_terminated_length": 206.55807647705078, "completions/min_length": 40.1, "completions/min_terminated_length": 40.1, "entropy": 0.028658897103741764, "epoch": 36.50015304560759, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.021240234375, "learning_rate": 5e-06, "loss": 0.0017, "num_tokens": 6106214047.0, "reward": 7.904139375686645, "reward_std": 0.4451727032661438, "rewards/combined_reward/mean": 7.904139280319214, "rewards/combined_reward/std": 2.3312257289886475, "step": 59640 }, { "clip_ratio/high_max": 0.0025482299650320785, "clip_ratio/high_mean": 0.0015137705954354412, "clip_ratio/low_mean": 0.0010755928118063214, "clip_ratio/low_min": 0.0003264401651298006, "clip_ratio/region_mean": 0.0025893634039675815, "completions/clipped_ratio": 0.0, "completions/max_length": 527.9, "completions/max_terminated_length": 527.9, "completions/mean_length": 206.24844207763672, "completions/mean_terminated_length": 206.24844207763672, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "entropy": 0.02906958165888985, "epoch": 36.51851851851852, "frac_reward_zero_std": 0.14166667088866233, "grad_norm": 0.01300048828125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6109306201.0, "reward": 8.03647198677063, "reward_std": 0.34893584847450254, "rewards/combined_reward/mean": 8.036471700668335, "rewards/combined_reward/std": 2.109134590625763, "step": 59670 }, { "clip_ratio/high_max": 0.0024620955780847, "clip_ratio/high_mean": 0.001380496292282866, "clip_ratio/low_mean": 0.0009075022486892218, "clip_ratio/low_min": 0.00024657738880099107, "clip_ratio/region_mean": 0.0022879985216908002, "completions/clipped_ratio": 0.0, "completions/max_length": 543.7, "completions/max_terminated_length": 543.7, "completions/mean_length": 201.79479522705077, "completions/mean_terminated_length": 201.79479522705077, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.026511059693681695, "epoch": 36.53688399142945, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.0045166015625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6112381253.0, "reward": 7.777392911911011, "reward_std": 0.34249784648418424, "rewards/combined_reward/mean": 7.777392625808716, "rewards/combined_reward/std": 2.6625423550605776, "step": 59700 }, { "clip_ratio/high_max": 0.002640661325616141, "clip_ratio/high_mean": 0.0015758186978928278, "clip_ratio/low_mean": 0.0010208365483170686, "clip_ratio/low_min": 0.0003021347423782572, "clip_ratio/region_mean": 0.0025966552280200024, "completions/clipped_ratio": 0.0, "completions/max_length": 543.1, "completions/max_terminated_length": 543.1, "completions/mean_length": 200.86458740234374, "completions/mean_terminated_length": 200.86458740234374, "completions/min_length": 42.1, "completions/min_terminated_length": 42.1, "entropy": 0.028226420236751436, "epoch": 36.55524946434037, "frac_reward_zero_std": 0.17500000558793544, "grad_norm": 0.0191650390625, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 6115452733.0, "reward": 7.957239580154419, "reward_std": 0.41007896363735197, "rewards/combined_reward/mean": 7.957239198684692, "rewards/combined_reward/std": 2.346678364276886, "step": 59730 }, { "clip_ratio/high_max": 0.0025036212740815245, "clip_ratio/high_mean": 0.001504322376907415, "clip_ratio/low_mean": 0.0010538146825031921, "clip_ratio/low_min": 0.00036203437969864656, "clip_ratio/region_mean": 0.0025581370602594687, "completions/clipped_ratio": 0.0, "completions/max_length": 541.5, "completions/max_terminated_length": 541.5, "completions/mean_length": 202.9322967529297, "completions/mean_terminated_length": 202.9322967529297, "completions/min_length": 41.7, "completions/min_terminated_length": 41.7, "entropy": 0.02981232775685688, "epoch": 36.5736149372513, "frac_reward_zero_std": 0.17083333618938923, "grad_norm": 0.008056640625, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6118532153.0, "reward": 7.996662425994873, "reward_std": 0.3091412991285324, "rewards/combined_reward/mean": 7.996662235260009, "rewards/combined_reward/std": 2.204971206188202, "step": 59760 }, { "clip_ratio/high_max": 0.0023021120126941242, "clip_ratio/high_mean": 0.0012785301666250841, "clip_ratio/low_mean": 0.0010068680502930269, "clip_ratio/low_min": 0.00028970979255973364, "clip_ratio/region_mean": 0.0022853982363206645, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 195.58932800292968, "completions/mean_terminated_length": 195.58932800292968, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "entropy": 0.0284834961562107, "epoch": 36.59198041016223, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.021728515625, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 6121583376.0, "reward": 8.079510736465455, "reward_std": 0.29728585928678514, "rewards/combined_reward/mean": 8.079510593414307, "rewards/combined_reward/std": 2.292326867580414, "step": 59790 }, { "clip_ratio/high_max": 0.002476133476496519, "clip_ratio/high_mean": 0.0013921278650135113, "clip_ratio/low_mean": 0.0009304244017888171, "clip_ratio/low_min": 0.00023278455725327755, "clip_ratio/region_mean": 0.0023225522595263707, "completions/clipped_ratio": 0.0, "completions/max_length": 606.3, "completions/max_terminated_length": 606.3, "completions/mean_length": 204.19479675292968, "completions/mean_terminated_length": 204.19479675292968, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.028148121697207294, "epoch": 36.61034588307316, "frac_reward_zero_std": 0.1458333373069763, "grad_norm": 0.0093994140625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6124667644.0, "reward": 8.179471731185913, "reward_std": 0.3992107167840004, "rewards/combined_reward/mean": 8.17947154045105, "rewards/combined_reward/std": 2.1314127683639525, "step": 59820 }, { "clip_ratio/high_max": 0.0023531577637186272, "clip_ratio/high_mean": 0.0012648025327507638, "clip_ratio/low_mean": 0.0008725522469831048, "clip_ratio/low_min": 0.00022193955956026912, "clip_ratio/region_mean": 0.002137354759906884, "completions/clipped_ratio": 0.0, "completions/max_length": 549.5, "completions/max_terminated_length": 549.5, "completions/mean_length": 197.52448425292968, "completions/mean_terminated_length": 197.52448425292968, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.02663563157742222, "epoch": 36.62871135598408, "frac_reward_zero_std": 0.20000000819563865, "grad_norm": 0.01165771484375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6127726298.0, "reward": 7.6519733428955075, "reward_std": 0.33732796311378477, "rewards/combined_reward/mean": 7.651973056793213, "rewards/combined_reward/std": 2.671976637840271, "step": 59850 }, { "clip_ratio/high_max": 0.002823774236215589, "clip_ratio/high_mean": 0.0015478630259167403, "clip_ratio/low_mean": 0.0009429339545022231, "clip_ratio/low_min": 0.00025075366866076363, "clip_ratio/region_mean": 0.0024907970223769856, "completions/clipped_ratio": 0.0, "completions/max_length": 553.6, "completions/max_terminated_length": 553.6, "completions/mean_length": 185.19844512939454, "completions/mean_terminated_length": 185.19844512939454, "completions/min_length": 38.4, "completions/min_terminated_length": 38.4, "entropy": 0.030105637138088545, "epoch": 36.64707682889501, "frac_reward_zero_std": 0.15833333916962147, "grad_norm": 0.01123046875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6130737620.0, "reward": 8.135231733322144, "reward_std": 0.29975809454917907, "rewards/combined_reward/mean": 8.135231447219848, "rewards/combined_reward/std": 2.1312952876091003, "step": 59880 }, { "clip_ratio/high_max": 0.002302513483785636, "clip_ratio/high_mean": 0.0013095693843448923, "clip_ratio/low_mean": 0.0010006832765308597, "clip_ratio/low_min": 0.0003029283885553014, "clip_ratio/region_mean": 0.0023102526747000715, "completions/clipped_ratio": 0.0, "completions/max_length": 572.5, "completions/max_terminated_length": 572.5, "completions/mean_length": 189.87734680175782, "completions/mean_terminated_length": 189.87734680175782, "completions/min_length": 39.1, "completions/min_terminated_length": 39.1, "entropy": 0.02512358423167219, "epoch": 36.66544230180594, "frac_reward_zero_std": 0.20833334028720857, "grad_norm": 0.00811767578125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6133766909.0, "reward": 7.8985474586486815, "reward_std": 0.3148704171180725, "rewards/combined_reward/mean": 7.898547172546387, "rewards/combined_reward/std": 2.3133850574493406, "step": 59910 }, { "clip_ratio/high_max": 0.0025906919152475895, "clip_ratio/high_mean": 0.0014436703437240793, "clip_ratio/low_mean": 0.0010086258600495058, "clip_ratio/low_min": 0.0002426387768840262, "clip_ratio/region_mean": 0.002452296201954596, "completions/clipped_ratio": 0.0, "completions/max_length": 513.6, "completions/max_terminated_length": 513.6, "completions/mean_length": 199.21614990234374, "completions/mean_terminated_length": 199.21614990234374, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.026558532286435364, "epoch": 36.68380777471687, "frac_reward_zero_std": 0.13750000558793546, "grad_norm": 0.0026702880859375, "learning_rate": 5e-06, "loss": 0.0018, "num_tokens": 6136832059.0, "reward": 8.065501880645751, "reward_std": 0.3899938099086285, "rewards/combined_reward/mean": 8.065501737594605, "rewards/combined_reward/std": 2.194762575626373, "step": 59940 }, { "clip_ratio/high_max": 0.0024168081833825755, "clip_ratio/high_mean": 0.0013927358631917741, "clip_ratio/low_mean": 0.0009752567122632172, "clip_ratio/low_min": 0.0002592688967221572, "clip_ratio/region_mean": 0.0023679925914620982, "completions/clipped_ratio": 0.0, "completions/max_length": 529.2, "completions/max_terminated_length": 529.2, "completions/mean_length": 194.2984420776367, "completions/mean_terminated_length": 194.2984420776367, "completions/min_length": 42.3, "completions/min_terminated_length": 42.3, "entropy": 0.028714586483935513, "epoch": 36.70217324762779, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0230712890625, "learning_rate": 5e-06, "loss": 0.0001, "num_tokens": 6139878325.0, "reward": 8.024220657348632, "reward_std": 0.3277119904756546, "rewards/combined_reward/mean": 8.024220371246338, "rewards/combined_reward/std": 2.20546156167984, "step": 59970 }, { "clip_ratio/high_max": 0.0024832503433572127, "clip_ratio/high_mean": 0.0014453509325297394, "clip_ratio/low_mean": 0.0011017890664030952, "clip_ratio/low_min": 0.00030534418789708676, "clip_ratio/region_mean": 0.002547139979530281, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 189.4546905517578, "completions/mean_terminated_length": 189.4546905517578, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.02782174399277816, "epoch": 36.72053872053872, "frac_reward_zero_std": 0.19583334028720856, "grad_norm": 0.0167236328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6142905991.0, "reward": 7.835037899017334, "reward_std": 0.33315608352422715, "rewards/combined_reward/mean": 7.8350378513336185, "rewards/combined_reward/std": 2.4791616916656496, "step": 60000 }, { "clip_ratio/high_max": 0.0022010034299455583, "clip_ratio/high_mean": 0.0013381298493186477, "clip_ratio/low_mean": 0.0007587791828579308, "clip_ratio/low_min": 0.00020116914699125726, "clip_ratio/region_mean": 0.0020969090323584775, "completions/clipped_ratio": 0.0, "completions/max_length": 560.6, "completions/max_terminated_length": 560.6, "completions/mean_length": 209.74714202880858, "completions/mean_terminated_length": 209.74714202880858, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "entropy": 0.031748458122213684, "epoch": 36.73829201101928, "frac_reward_zero_std": 0.1625000074505806, "grad_norm": 0.01220703125, "learning_rate": 5e-06, "loss": 0.0019, "num_tokens": 6146011580.0, "reward": 7.995605564117431, "reward_std": 0.3649419084191322, "rewards/combined_reward/mean": 7.995605039596557, "rewards/combined_reward/std": 2.5295765519142153, "step": 60030 }, { "clip_ratio/high_max": 0.0018103700858773664, "clip_ratio/high_mean": 0.0009832295242328353, "clip_ratio/low_mean": 0.0008034945063021344, "clip_ratio/low_min": 0.0001932421687039702, "clip_ratio/region_mean": 0.00178672402932231, "completions/clipped_ratio": 0.0, "completions/max_length": 613.2, "completions/max_terminated_length": 613.2, "completions/mean_length": 202.10833587646485, "completions/mean_terminated_length": 202.10833587646485, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.026359116937965155, "epoch": 36.75665748393021, "frac_reward_zero_std": 0.1541666690260172, "grad_norm": 0.01422119140625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6149087836.0, "reward": 8.17109236717224, "reward_std": 0.34614808708429334, "rewards/combined_reward/mean": 8.171091842651368, "rewards/combined_reward/std": 2.017248731851578, "step": 60060 }, { "clip_ratio/high_max": 0.001988391222403152, "clip_ratio/high_mean": 0.001162893826403888, "clip_ratio/low_mean": 0.0007216581701868563, "clip_ratio/low_min": 0.00021061977677163667, "clip_ratio/region_mean": 0.0018845520132648138, "completions/clipped_ratio": 0.0, "completions/max_length": 570.4, "completions/max_terminated_length": 570.4, "completions/mean_length": 211.8257858276367, "completions/mean_terminated_length": 211.8257858276367, "completions/min_length": 43.7, "completions/min_terminated_length": 43.7, "entropy": 0.027453064902996024, "epoch": 36.77502295684114, "frac_reward_zero_std": 0.1708333373069763, "grad_norm": 0.0074462890625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6152201407.0, "reward": 8.142016649246216, "reward_std": 0.32791845351457594, "rewards/combined_reward/mean": 8.14201636314392, "rewards/combined_reward/std": 2.300868773460388, "step": 60090 }, { "clip_ratio/high_max": 0.002044012509092378, "clip_ratio/high_mean": 0.0011315541799073495, "clip_ratio/low_mean": 0.0007198089245018006, "clip_ratio/low_min": 0.00020285683179584643, "clip_ratio/region_mean": 0.0018513631201737249, "completions/clipped_ratio": 0.0, "completions/max_length": 596.9, "completions/max_terminated_length": 596.9, "completions/mean_length": 208.7945358276367, "completions/mean_terminated_length": 208.7945358276367, "completions/min_length": 45.7, "completions/min_terminated_length": 45.7, "entropy": 0.02821128289991369, "epoch": 36.79338842975206, "frac_reward_zero_std": 0.16250000335276127, "grad_norm": 0.01171875, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6155303338.0, "reward": 8.157721662521363, "reward_std": 0.2521477028727531, "rewards/combined_reward/mean": 8.157721376419067, "rewards/combined_reward/std": 2.0745994448661804, "step": 60120 }, { "clip_ratio/high_max": 0.002204254914734823, "clip_ratio/high_mean": 0.0012024933838498933, "clip_ratio/low_mean": 0.0007397785186185502, "clip_ratio/low_min": 0.000137029233398304, "clip_ratio/region_mean": 0.0019422718784577835, "completions/clipped_ratio": 0.0, "completions/max_length": 580.2, "completions/max_terminated_length": 580.2, "completions/mean_length": 199.91693115234375, "completions/mean_terminated_length": 199.91693115234375, "completions/min_length": 39.7, "completions/min_terminated_length": 39.7, "entropy": 0.03053381498903036, "epoch": 36.81175390266299, "frac_reward_zero_std": 0.21250000670552255, "grad_norm": 0.010986328125, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6158371179.0, "reward": 7.9497462749481205, "reward_std": 0.30010201036930084, "rewards/combined_reward/mean": 7.949746131896973, "rewards/combined_reward/std": 2.2839062452316283, "step": 60150 }, { "clip_ratio/high_max": 0.002263490053398224, "clip_ratio/high_mean": 0.0012733247252375197, "clip_ratio/low_mean": 0.0008278546879106822, "clip_ratio/low_min": 0.00027286618836418105, "clip_ratio/region_mean": 0.002101179400536542, "completions/clipped_ratio": 0.0, "completions/max_length": 565.5, "completions/max_terminated_length": 565.5, "completions/mean_length": 203.93698425292968, "completions/mean_terminated_length": 203.93698425292968, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.028283760789781807, "epoch": 36.83011937557392, "frac_reward_zero_std": 0.21250000596046448, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6161454457.0, "reward": 8.142348527908325, "reward_std": 0.276405468583107, "rewards/combined_reward/mean": 8.142347955703736, "rewards/combined_reward/std": 2.3293092370033266, "step": 60180 }, { "clip_ratio/high_max": 0.0031618790710732964, "clip_ratio/high_mean": 0.0016652902209898456, "clip_ratio/low_mean": 0.0007402347391083215, "clip_ratio/low_min": 0.00021349823485555436, "clip_ratio/region_mean": 0.002405524970769572, "completions/clipped_ratio": 0.0, "completions/max_length": 603.4, "completions/max_terminated_length": 603.4, "completions/mean_length": 201.64193115234374, "completions/mean_terminated_length": 201.64193115234374, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.033225487809007366, "epoch": 36.84848484848485, "frac_reward_zero_std": 0.19166667237877846, "grad_norm": 0.0135498046875, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6164528922.0, "reward": 8.274881029129029, "reward_std": 0.3304283797740936, "rewards/combined_reward/mean": 8.274881029129029, "rewards/combined_reward/std": 2.0586212277412415, "step": 60210 }, { "clip_ratio/high_max": 0.001943459413208378, "clip_ratio/high_mean": 0.0010855247632813795, "clip_ratio/low_mean": 0.0009552978364808951, "clip_ratio/low_min": 0.00023522258901114885, "clip_ratio/region_mean": 0.0020408225842402316, "completions/clipped_ratio": 0.0, "completions/max_length": 619.3, "completions/max_terminated_length": 619.3, "completions/mean_length": 204.2984420776367, "completions/mean_terminated_length": 204.2984420776367, "completions/min_length": 53.8, "completions/min_terminated_length": 53.8, "entropy": 0.030482247401960195, "epoch": 36.86685032139577, "frac_reward_zero_std": 0.1625000037252903, "grad_norm": 0.035400390625, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6167613588.0, "reward": 8.290172672271728, "reward_std": 0.26410685628652575, "rewards/combined_reward/mean": 8.290172624588013, "rewards/combined_reward/std": 1.9387029469013215, "step": 60240 }, { "clip_ratio/high_max": 0.001878358605123746, "clip_ratio/high_mean": 0.0011047641307717034, "clip_ratio/low_mean": 0.0007545156507452097, "clip_ratio/low_min": 0.00023266172186898378, "clip_ratio/region_mean": 0.001859279757384987, "completions/clipped_ratio": 0.0, "completions/max_length": 613.5, "completions/max_terminated_length": 613.5, "completions/mean_length": 210.50234985351562, "completions/mean_terminated_length": 210.50234985351562, "completions/min_length": 40.9, "completions/min_terminated_length": 40.9, "entropy": 0.02783806392302116, "epoch": 36.8852157943067, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.0078125, "learning_rate": 5e-06, "loss": 0.0034, "num_tokens": 6170722077.0, "reward": 8.27436466217041, "reward_std": 0.3628034755587578, "rewards/combined_reward/mean": 8.274364328384399, "rewards/combined_reward/std": 2.0409519195556642, "step": 60270 }, { "clip_ratio/high_max": 0.001549404425895773, "clip_ratio/high_mean": 0.0009079823561478406, "clip_ratio/low_mean": 0.0007453262913865425, "clip_ratio/low_min": 0.00013990464431117288, "clip_ratio/region_mean": 0.001653308650808564, "completions/clipped_ratio": 0.0, "completions/max_length": 625.9, "completions/max_terminated_length": 625.9, "completions/mean_length": 214.01797637939453, "completions/mean_terminated_length": 214.01797637939453, "completions/min_length": 46.8, "completions/min_terminated_length": 46.8, "entropy": 0.0259612287239482, "epoch": 36.90358126721763, "frac_reward_zero_std": 0.13333333805203437, "grad_norm": 0.01275634765625, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6173844066.0, "reward": 7.986669969558716, "reward_std": 0.3492489606142044, "rewards/combined_reward/mean": 7.986669635772705, "rewards/combined_reward/std": 2.1417524456977843, "step": 60300 }, { "clip_ratio/high_max": 0.0024454933217687843, "clip_ratio/high_mean": 0.001344558141863672, "clip_ratio/low_mean": 0.0008564603244546258, "clip_ratio/low_min": 0.00022414449388937405, "clip_ratio/region_mean": 0.0022010184785661597, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 188.56068420410156, "completions/mean_terminated_length": 188.56068420410156, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "entropy": 0.03185348082333803, "epoch": 36.92194674012856, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.006561279296875, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 6176868299.0, "reward": 8.319356441497803, "reward_std": 0.2854164853692055, "rewards/combined_reward/mean": 8.319356346130371, "rewards/combined_reward/std": 1.8687144219875336, "step": 60330 }, { "clip_ratio/high_max": 0.001939525897129594, "clip_ratio/high_mean": 0.0010872551470432275, "clip_ratio/low_mean": 0.0007477245567618714, "clip_ratio/low_min": 0.00020438863139133901, "clip_ratio/region_mean": 0.0018349797042901627, "completions/clipped_ratio": 0.0, "completions/max_length": 678.9, "completions/max_terminated_length": 678.9, "completions/mean_length": 203.89349517822265, "completions/mean_terminated_length": 203.89349517822265, "completions/min_length": 45.8, "completions/min_terminated_length": 45.8, "entropy": 0.027989901493613918, "epoch": 36.94031221303948, "frac_reward_zero_std": 0.2041666753590107, "grad_norm": 0.01190185546875, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 6179951410.0, "reward": 7.940687751770019, "reward_std": 0.26497893780469894, "rewards/combined_reward/mean": 7.940687561035157, "rewards/combined_reward/std": 2.556736469268799, "step": 60360 }, { "clip_ratio/high_max": 0.0016252480648593821, "clip_ratio/high_mean": 0.000908645930030616, "clip_ratio/low_mean": 0.0006452756473057283, "clip_ratio/low_min": 0.00018324590419069864, "clip_ratio/region_mean": 0.0015539215683626631, "completions/clipped_ratio": 0.0, "completions/max_length": 581.5, "completions/max_terminated_length": 581.5, "completions/mean_length": 205.97369995117188, "completions/mean_terminated_length": 205.97369995117188, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 0.025537174039830763, "epoch": 36.95867768595041, "frac_reward_zero_std": 0.1833333380520344, "grad_norm": 0.0040283203125, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6183042509.0, "reward": 7.945443487167358, "reward_std": 0.2535731166601181, "rewards/combined_reward/mean": 7.945443201065063, "rewards/combined_reward/std": 2.290303874015808, "step": 60390 }, { "clip_ratio/high_max": 0.0017500985200361659, "clip_ratio/high_mean": 0.0010657495532844527, "clip_ratio/low_mean": 0.0006302042354946024, "clip_ratio/low_min": 0.00013113425763246293, "clip_ratio/region_mean": 0.0016959537970251404, "completions/clipped_ratio": 0.0, "completions/max_length": 585.1, "completions/max_terminated_length": 585.1, "completions/mean_length": 199.27838897705078, "completions/mean_terminated_length": 199.27838897705078, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "entropy": 0.025884551849837106, "epoch": 36.97704315886134, "frac_reward_zero_std": 0.2041666716337204, "grad_norm": 0.01129150390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6186107898.0, "reward": 8.085112476348877, "reward_std": 0.3490646734833717, "rewards/combined_reward/mean": 8.085112142562867, "rewards/combined_reward/std": 2.2755405724048616, "step": 60420 }, { "clip_ratio/high_max": 0.0018769755508401432, "clip_ratio/high_mean": 0.0010301382453083837, "clip_ratio/low_mean": 0.0008453347293349604, "clip_ratio/low_min": 0.0002881676065347468, "clip_ratio/region_mean": 0.001875472977796259, "completions/clipped_ratio": 0.0, "completions/max_length": 563.3, "completions/max_terminated_length": 563.3, "completions/mean_length": 201.74062805175782, "completions/mean_terminated_length": 201.74062805175782, "completions/min_length": 44.9, "completions/min_terminated_length": 44.9, "entropy": 0.025843802296246093, "epoch": 36.99540863177227, "frac_reward_zero_std": 0.1458333358168602, "grad_norm": 0.005859375, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6189182742.0, "reward": 8.118364906311035, "reward_std": 0.29999661445617676, "rewards/combined_reward/mean": 8.118364953994751, "rewards/combined_reward/std": 2.206037366390228, "step": 60450 }, { "clip_ratio/high_max": 0.002311651210861708, "clip_ratio/high_mean": 0.001333564767145483, "clip_ratio/low_mean": 0.0008444937975284239, "clip_ratio/low_min": 0.00020562158157995552, "clip_ratio/region_mean": 0.002178058555906974, "completions/clipped_ratio": 0.0, "completions/max_length": 544.8181818181819, "completions/max_terminated_length": 544.8181818181819, "completions/mean_length": 205.00876409357244, "completions/mean_terminated_length": 205.00876409357244, "completions/min_length": 43.72727272727273, "completions/min_terminated_length": 43.72727272727273, "entropy": 0.02904677294866472, "epoch": 37.01408019589838, "frac_reward_zero_std": 0.18560606850819153, "grad_norm": 0.01239013671875, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 6192578875.0, "reward": 8.178891311992299, "reward_std": 0.3200361186807806, "rewards/combined_reward/mean": 8.178891225294633, "rewards/combined_reward/std": 2.128684650767933, "step": 60480 }, { "clip_ratio/high_max": 0.002491404055763269, "clip_ratio/high_mean": 0.0015103426349620955, "clip_ratio/low_mean": 0.0009876768547352791, "clip_ratio/low_min": 0.0002991968414183551, "clip_ratio/region_mean": 0.0024980194890910448, "completions/clipped_ratio": 0.0, "completions/max_length": 537.4, "completions/max_terminated_length": 537.4, "completions/mean_length": 199.7414123535156, "completions/mean_terminated_length": 199.7414123535156, "completions/min_length": 37.1, "completions/min_terminated_length": 37.1, "entropy": 0.030614282959140836, "epoch": 37.0324456688093, "frac_reward_zero_std": 0.16666666977107525, "grad_norm": 0.0257568359375, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6195646042.0, "reward": 7.92452220916748, "reward_std": 0.34527396485209466, "rewards/combined_reward/mean": 7.924522018432617, "rewards/combined_reward/std": 2.4852097392082215, "step": 60510 }, { "clip_ratio/high_max": 0.003196258085760443, "clip_ratio/high_mean": 0.0019278426167147699, "clip_ratio/low_mean": 0.000986052709413343, "clip_ratio/low_min": 0.00023785665155931686, "clip_ratio/region_mean": 0.002913895314850379, "completions/clipped_ratio": 0.0, "completions/max_length": 583.4, "completions/max_terminated_length": 583.4, "completions/mean_length": 211.39505767822266, "completions/mean_terminated_length": 211.39505767822266, "completions/min_length": 43.1, "completions/min_terminated_length": 43.1, "entropy": 0.03259018369329472, "epoch": 37.05081114172023, "frac_reward_zero_std": 0.14166667275130748, "grad_norm": 0.0137939453125, "learning_rate": 5e-06, "loss": 0.0008, "num_tokens": 6198757959.0, "reward": 8.202747106552124, "reward_std": 0.3024443775415421, "rewards/combined_reward/mean": 8.202746486663818, "rewards/combined_reward/std": 2.000635004043579, "step": 60540 }, { "clip_ratio/high_max": 0.0020694057234019662, "clip_ratio/high_mean": 0.0011426436066055126, "clip_ratio/low_mean": 0.0008382113535238508, "clip_ratio/low_min": 0.00013737305416725576, "clip_ratio/region_mean": 0.0019808549564913846, "completions/clipped_ratio": 0.0, "completions/max_length": 503.1, "completions/max_terminated_length": 503.1, "completions/mean_length": 198.6992218017578, "completions/mean_terminated_length": 198.6992218017578, "completions/min_length": 38.2, "completions/min_terminated_length": 38.2, "entropy": 0.024822839706515273, "epoch": 37.06917661463116, "frac_reward_zero_std": 0.11666666977107525, "grad_norm": 0.008544921875, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6201821124.0, "reward": 8.317335224151611, "reward_std": 0.3208719491958618, "rewards/combined_reward/mean": 8.317334938049317, "rewards/combined_reward/std": 2.097157371044159, "step": 60570 }, { "clip_ratio/high_max": 0.002743625883401061, "clip_ratio/high_mean": 0.0016020483989753606, "clip_ratio/low_mean": 0.000993663681704978, "clip_ratio/low_min": 0.0002625178928913859, "clip_ratio/region_mean": 0.0025957120922006044, "completions/clipped_ratio": 0.0, "completions/max_length": 618.8, "completions/max_terminated_length": 618.8, "completions/mean_length": 202.7619857788086, "completions/mean_terminated_length": 202.7619857788086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.0317182814857612, "epoch": 37.08754208754209, "frac_reward_zero_std": 0.13750000447034835, "grad_norm": 0.009521484375, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 6204899890.0, "reward": 8.093045377731324, "reward_std": 0.28341620191931727, "rewards/combined_reward/mean": 8.09304518699646, "rewards/combined_reward/std": 2.091159462928772, "step": 60600 }, { "clip_ratio/high_max": 0.002183143353856091, "clip_ratio/high_mean": 0.001314189817821898, "clip_ratio/low_mean": 0.0009640878946811426, "clip_ratio/low_min": 0.00027318856179287345, "clip_ratio/region_mean": 0.002278277720688493, "completions/clipped_ratio": 0.0, "completions/max_length": 744.7, "completions/max_terminated_length": 744.7, "completions/mean_length": 209.2713607788086, "completions/mean_terminated_length": 209.2713607788086, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "entropy": 0.028659445388863485, "epoch": 37.10590756045301, "frac_reward_zero_std": 0.17916667386889457, "grad_norm": 0.00885009765625, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6208003652.0, "reward": 7.921809816360474, "reward_std": 0.3146898984909058, "rewards/combined_reward/mean": 7.9218096256256105, "rewards/combined_reward/std": 2.2413519978523255, "step": 60630 }, { "clip_ratio/high_max": 0.002289204165087237, "clip_ratio/high_mean": 0.0013311001409116823, "clip_ratio/low_mean": 0.0009657165363023523, "clip_ratio/low_min": 0.0002344812889835642, "clip_ratio/region_mean": 0.0022968166895831624, "completions/clipped_ratio": 0.0, "completions/max_length": 612.4, "completions/max_terminated_length": 612.4, "completions/mean_length": 204.06979522705078, "completions/mean_terminated_length": 204.06979522705078, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 0.02998416922055185, "epoch": 37.12427303336394, "frac_reward_zero_std": 0.16666667014360428, "grad_norm": 0.0107421875, "learning_rate": 5e-06, "loss": 0.0016, "num_tokens": 6211087440.0, "reward": 7.989088630676269, "reward_std": 0.26193844228982927, "rewards/combined_reward/mean": 7.989088296890259, "rewards/combined_reward/std": 2.463514268398285, "step": 60660 }, { "clip_ratio/high_max": 0.002420163661008701, "clip_ratio/high_mean": 0.0013477934085434147, "clip_ratio/low_mean": 0.0010300867831877743, "clip_ratio/low_min": 0.0002922490697528701, "clip_ratio/region_mean": 0.002377880195854232, "completions/clipped_ratio": 0.0, "completions/max_length": 645.2, "completions/max_terminated_length": 645.2, "completions/mean_length": 197.35937957763673, "completions/mean_terminated_length": 197.35937957763673, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.029481424655144414, "epoch": 37.14263850627487, "frac_reward_zero_std": 0.21666667461395264, "grad_norm": 0.02099609375, "learning_rate": 5e-06, "loss": 0.002, "num_tokens": 6214145460.0, "reward": 7.931514263153076, "reward_std": 0.31033128723502157, "rewards/combined_reward/mean": 7.931514024734497, "rewards/combined_reward/std": 2.3295389890670775, "step": 60690 }, { "clip_ratio/high_max": 0.0021574515349736126, "clip_ratio/high_mean": 0.0012509938390091217, "clip_ratio/low_mean": 0.0008524422287640239, "clip_ratio/low_min": 0.00020710518389629822, "clip_ratio/region_mean": 0.002103436045096411, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 203.80964202880858, "completions/mean_terminated_length": 203.80964202880858, "completions/min_length": 43.5, "completions/min_terminated_length": 43.5, "entropy": 0.02780179106630385, "epoch": 37.1610039791858, "frac_reward_zero_std": 0.13333333618938922, "grad_norm": 0.0091552734375, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6217228249.0, "reward": 8.153032732009887, "reward_std": 0.320121967792511, "rewards/combined_reward/mean": 8.153032732009887, "rewards/combined_reward/std": 2.108227479457855, "step": 60720 }, { "clip_ratio/high_max": 0.0022883495514785562, "clip_ratio/high_mean": 0.0012892773656252151, "clip_ratio/low_mean": 0.0009230855369726972, "clip_ratio/low_min": 0.00027953182749721844, "clip_ratio/region_mean": 0.0022123629110865294, "completions/clipped_ratio": 0.0, "completions/max_length": 563.1, "completions/max_terminated_length": 563.1, "completions/mean_length": 202.29662170410157, "completions/mean_terminated_length": 202.29662170410157, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "entropy": 0.025763744278810917, "epoch": 37.17936945209672, "frac_reward_zero_std": 0.1666666731238365, "grad_norm": 0.01141357421875, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 6220305228.0, "reward": 8.053554725646972, "reward_std": 0.2941873177886009, "rewards/combined_reward/mean": 8.053554487228393, "rewards/combined_reward/std": 2.3774508059024813, "step": 60750 }, { "clip_ratio/high_max": 0.002515348729987939, "clip_ratio/high_mean": 0.0013955170720388803, "clip_ratio/low_mean": 0.0008761482184733419, "clip_ratio/low_min": 0.00019483079619628068, "clip_ratio/region_mean": 0.0022716652708671364, "completions/clipped_ratio": 0.0, "completions/max_length": 536.3, "completions/max_terminated_length": 536.3, "completions/mean_length": 195.51849365234375, "completions/mean_terminated_length": 195.51849365234375, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "entropy": 0.027591845338853695, "epoch": 37.19773492500765, "frac_reward_zero_std": 0.1916666727513075, "grad_norm": 0.01068115234375, "learning_rate": 5e-06, "loss": 0.0009, "num_tokens": 6223356179.0, "reward": 8.161541223526001, "reward_std": 0.2141267403960228, "rewards/combined_reward/mean": 8.161540937423705, "rewards/combined_reward/std": 2.0967944264411926, "step": 60780 }, { "clip_ratio/high_max": 0.0024692296584059173, "clip_ratio/high_mean": 0.0014496783580398187, "clip_ratio/low_mean": 0.0009499037674686406, "clip_ratio/low_min": 0.00025655494246166197, "clip_ratio/region_mean": 0.0023995821247808635, "completions/clipped_ratio": 0.0, "completions/max_length": 657.2, "completions/max_terminated_length": 657.2, "completions/mean_length": 200.47995147705078, "completions/mean_terminated_length": 200.47995147705078, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.028477749686377743, "epoch": 37.21610039791858, "frac_reward_zero_std": 0.15416667126119138, "grad_norm": 0.0186767578125, "learning_rate": 5e-06, "loss": 0.0006, "num_tokens": 6226426182.0, "reward": 8.02132158279419, "reward_std": 0.32127553075551984, "rewards/combined_reward/mean": 8.021321010589599, "rewards/combined_reward/std": 2.473380517959595, "step": 60810 }, { "clip_ratio/high_max": 0.002745241978360961, "clip_ratio/high_mean": 0.001575590460076152, "clip_ratio/low_mean": 0.0010332490537014868, "clip_ratio/low_min": 0.000265122638666071, "clip_ratio/region_mean": 0.0026088395068654792, "completions/clipped_ratio": 0.0, "completions/max_length": 613.3, "completions/max_terminated_length": 613.3, "completions/mean_length": 198.4182342529297, "completions/mean_terminated_length": 198.4182342529297, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "entropy": 0.03496701999101788, "epoch": 37.23446587082951, "frac_reward_zero_std": 0.16666667237877847, "grad_norm": 0.0218505859375, "learning_rate": 5e-06, "loss": 0.001, "num_tokens": 6229488268.0, "reward": 8.076002025604248, "reward_std": 0.2591355934739113, "rewards/combined_reward/mean": 8.076001691818238, "rewards/combined_reward/std": 2.1072788715362547, "step": 60840 }, { "clip_ratio/high_max": 0.002007974367006682, "clip_ratio/high_mean": 0.0011358440886397147, "clip_ratio/low_mean": 0.0007279336953312546, "clip_ratio/low_min": 0.00013718186164624058, "clip_ratio/region_mean": 0.0018637777960975655, "completions/clipped_ratio": 0.0, "completions/max_length": 620.6, "completions/max_terminated_length": 620.6, "completions/mean_length": 201.0796920776367, "completions/mean_terminated_length": 201.0796920776367, "completions/min_length": 37.9, "completions/min_terminated_length": 37.9, "entropy": 0.023838228918612002, "epoch": 37.25283134374043, "frac_reward_zero_std": 0.16250000521540642, "grad_norm": 0.0263671875, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6232560574.0, "reward": 8.126776790618896, "reward_std": 0.31179494708776473, "rewards/combined_reward/mean": 8.126776695251465, "rewards/combined_reward/std": 2.21409353017807, "step": 60870 }, { "clip_ratio/high_max": 0.002619988396569776, "clip_ratio/high_mean": 0.0015601963828279016, "clip_ratio/low_mean": 0.0010429584409090845, "clip_ratio/low_min": 0.00041977497263966747, "clip_ratio/region_mean": 0.00260315482931522, "completions/clipped_ratio": 0.0, "completions/max_length": 494.7, "completions/max_terminated_length": 494.7, "completions/mean_length": 194.8312545776367, "completions/mean_terminated_length": 194.8312545776367, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.03061138146246473, "epoch": 37.27119681665136, "frac_reward_zero_std": 0.2333333410322666, "grad_norm": 0.0169677734375, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6235608886.0, "reward": 8.219816970825196, "reward_std": 0.24989786520600318, "rewards/combined_reward/mean": 8.219816780090332, "rewards/combined_reward/std": 2.10613032579422, "step": 60900 }, { "clip_ratio/high_max": 0.0020120821262632187, "clip_ratio/high_mean": 0.0010999322207984126, "clip_ratio/low_mean": 0.0008987811960347851, "clip_ratio/low_min": 0.00020209017408584867, "clip_ratio/region_mean": 0.001998713412710155, "completions/clipped_ratio": 0.0, "completions/max_length": 653.5, "completions/max_terminated_length": 653.5, "completions/mean_length": 199.0565170288086, "completions/mean_terminated_length": 199.0565170288086, "completions/min_length": 37.8, "completions/min_terminated_length": 37.8, "entropy": 0.02395031265138338, "epoch": 37.28956228956229, "frac_reward_zero_std": 0.19166667349636554, "grad_norm": 0.01422119140625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6238673423.0, "reward": 7.853608560562134, "reward_std": 0.32776085287332535, "rewards/combined_reward/mean": 7.853608083724976, "rewards/combined_reward/std": 2.443291258811951, "step": 60930 }, { "clip_ratio/high_max": 0.0022598118036209294, "clip_ratio/high_mean": 0.001305660123640943, "clip_ratio/low_mean": 0.0009449845330285219, "clip_ratio/low_min": 0.00022551029057164365, "clip_ratio/region_mean": 0.002250644668432263, "completions/clipped_ratio": 0.0, "completions/max_length": 581.8, "completions/max_terminated_length": 581.8, "completions/mean_length": 200.18854675292968, "completions/mean_terminated_length": 200.18854675292968, "completions/min_length": 39.8, "completions/min_terminated_length": 39.8, "entropy": 0.027986601693555713, "epoch": 37.30792776247322, "frac_reward_zero_std": 0.1541666716337204, "grad_norm": 0.006378173828125, "learning_rate": 5e-06, "loss": 0.0014, "num_tokens": 6241742307.0, "reward": 8.115609264373779, "reward_std": 0.4030561849474907, "rewards/combined_reward/mean": 8.115609073638916, "rewards/combined_reward/std": 2.2109904050827027, "step": 60960 }, { "clip_ratio/high_max": 0.0024239847024243016, "clip_ratio/high_mean": 0.0014552370708164138, "clip_ratio/low_mean": 0.0008347061030993548, "clip_ratio/low_min": 0.00018088101254155238, "clip_ratio/region_mean": 0.0022899431904079394, "completions/clipped_ratio": 0.0, "completions/max_length": 564.8, "completions/max_terminated_length": 564.8, "completions/mean_length": 200.24505767822265, "completions/mean_terminated_length": 200.24505767822265, "completions/min_length": 38.1, "completions/min_terminated_length": 38.1, "entropy": 0.029304335856189332, "epoch": 37.32629323538414, "frac_reward_zero_std": 0.12916666716337205, "grad_norm": 0.00665283203125, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6244811408.0, "reward": 8.281531429290771, "reward_std": 0.3228962063789368, "rewards/combined_reward/mean": 8.281531238555909, "rewards/combined_reward/std": 2.013891804218292, "step": 60990 }, { "clip_ratio/high_max": 0.0021980140008963645, "clip_ratio/high_mean": 0.0012954238807045234, "clip_ratio/low_mean": 0.0010525096007768298, "clip_ratio/low_min": 0.0004054521421494428, "clip_ratio/region_mean": 0.0023479334646253846, "completions/clipped_ratio": 0.0, "completions/max_length": 606.3, "completions/max_terminated_length": 606.3, "completions/mean_length": 208.57136077880858, "completions/mean_terminated_length": 208.57136077880858, "completions/min_length": 45.2, "completions/min_terminated_length": 45.2, "entropy": 0.030217418540269138, "epoch": 37.34465870829507, "frac_reward_zero_std": 0.16666667200624943, "grad_norm": 0.0211181640625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6247912482.0, "reward": 7.787561511993408, "reward_std": 0.3611500233411789, "rewards/combined_reward/mean": 7.787561225891113, "rewards/combined_reward/std": 2.6838305950164796, "step": 61020 }, { "clip_ratio/high_max": 0.0024875823646046532, "clip_ratio/high_mean": 0.0014395555624408492, "clip_ratio/low_mean": 0.0008946362388693766, "clip_ratio/low_min": 0.0002891331840752779, "clip_ratio/region_mean": 0.0023341918150739122, "completions/clipped_ratio": 0.0, "completions/max_length": 607.1, "completions/max_terminated_length": 607.1, "completions/mean_length": 203.6945373535156, "completions/mean_terminated_length": 203.6945373535156, "completions/min_length": 39.9, "completions/min_terminated_length": 39.9, "entropy": 0.02883010815518598, "epoch": 37.363024181206, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.00750732421875, "learning_rate": 5e-06, "loss": 0.0013, "num_tokens": 6250994829.0, "reward": 8.324559831619263, "reward_std": 0.2983748361468315, "rewards/combined_reward/mean": 8.324559593200684, "rewards/combined_reward/std": 1.9922597050666808, "step": 61050 }, { "clip_ratio/high_max": 0.0024033125346856346, "clip_ratio/high_mean": 0.001324117518015555, "clip_ratio/low_mean": 0.0010001956507646051, "clip_ratio/low_min": 0.0003024156949929117, "clip_ratio/region_mean": 0.0023243131717511764, "completions/clipped_ratio": 0.0, "completions/max_length": 623.4, "completions/max_terminated_length": 623.4, "completions/mean_length": 203.6088607788086, "completions/mean_terminated_length": 203.6088607788086, "completions/min_length": 37.7, "completions/min_terminated_length": 37.7, "entropy": 0.030032550947119794, "epoch": 37.38138965411693, "frac_reward_zero_std": 0.1791666716337204, "grad_norm": 0.0181884765625, "learning_rate": 5e-06, "loss": 0.0003, "num_tokens": 6254076847.0, "reward": 7.98789234161377, "reward_std": 0.27229504138231275, "rewards/combined_reward/mean": 7.987891769409179, "rewards/combined_reward/std": 2.3558094263076783, "step": 61080 }, { "clip_ratio/high_max": 0.003103303570969729, "clip_ratio/high_mean": 0.0017769270163019732, "clip_ratio/low_mean": 0.0008856737885556261, "clip_ratio/low_min": 0.0002743628402337587, "clip_ratio/region_mean": 0.0026626008077679825, "completions/clipped_ratio": 0.0, "completions/max_length": 541.5, "completions/max_terminated_length": 541.5, "completions/mean_length": 206.74687957763672, "completions/mean_terminated_length": 206.74687957763672, "completions/min_length": 42.4, "completions/min_terminated_length": 42.4, "entropy": 0.032221240007008115, "epoch": 37.39975512702785, "frac_reward_zero_std": 0.16666667088866233, "grad_norm": 0.0206298828125, "learning_rate": 5e-06, "loss": 0.0007, "num_tokens": 6257170915.0, "reward": 7.78069920539856, "reward_std": 0.3371389165520668, "rewards/combined_reward/mean": 7.780698919296265, "rewards/combined_reward/std": 2.458569049835205, "step": 61110 }, { "clip_ratio/high_max": 0.0021652967043337412, "clip_ratio/high_mean": 0.001102926839303109, "clip_ratio/low_mean": 0.0007728601365670329, "clip_ratio/low_min": 0.00019431728214840404, "clip_ratio/region_mean": 0.0018757869613182266, "completions/clipped_ratio": 0.0, "completions/max_length": 575.9, "completions/max_terminated_length": 575.9, "completions/mean_length": 200.26458892822265, "completions/mean_terminated_length": 200.26458892822265, "completions/min_length": 42.5, "completions/min_terminated_length": 42.5, "entropy": 0.025164976747085652, "epoch": 37.41812059993878, "frac_reward_zero_std": 0.17083333805203438, "grad_norm": 0.023681640625, "learning_rate": 5e-06, "loss": 0.0015, "num_tokens": 6260240091.0, "reward": 7.922118902206421, "reward_std": 0.28479465916752816, "rewards/combined_reward/mean": 7.922118234634399, "rewards/combined_reward/std": 2.255589997768402, "step": 61140 }, { "clip_ratio/high_max": 0.002075849771790672, "clip_ratio/high_mean": 0.001211252905462364, "clip_ratio/low_mean": 0.0008880503993471696, "clip_ratio/low_min": 0.0002272774475083376, "clip_ratio/region_mean": 0.002099303302869278, "completions/clipped_ratio": 0.0, "completions/max_length": 636.5, "completions/max_terminated_length": 636.5, "completions/mean_length": 204.5143280029297, "completions/mean_terminated_length": 204.5143280029297, "completions/min_length": 46.9, "completions/min_terminated_length": 46.9, "entropy": 0.02669293942550818, "epoch": 37.43648607284971, "frac_reward_zero_std": 0.15833333656191825, "grad_norm": 0.006988525390625, "learning_rate": 5e-06, "loss": 0.0011, "num_tokens": 6263325586.0, "reward": 8.0181574344635, "reward_std": 0.25842341780662537, "rewards/combined_reward/mean": 8.018157148361206, "rewards/combined_reward/std": 2.483305811882019, "step": 61170 }, { "clip_ratio/high_max": 0.0022298081957463486, "clip_ratio/high_mean": 0.0013743522061607412, "clip_ratio/low_mean": 0.0008827971541904844, "clip_ratio/low_min": 0.00030618409922074835, "clip_ratio/region_mean": 0.0022571493672633855, "completions/clipped_ratio": 0.0, "completions/max_length": 597.3, "completions/max_terminated_length": 597.3, "completions/mean_length": 198.43594512939453, "completions/mean_terminated_length": 198.43594512939453, "completions/min_length": 40.5, "completions/min_terminated_length": 40.5, "entropy": 0.03139874883927405, "epoch": 37.45485154576064, "frac_reward_zero_std": 0.22916667237877847, "grad_norm": 0.006103515625, "learning_rate": 5e-06, "loss": 0.0012, "num_tokens": 6266387740.0, "reward": 8.197391748428345, "reward_std": 0.31101965606212617, "rewards/combined_reward/mean": 8.197391557693482, "rewards/combined_reward/std": 2.2144232034683227, "step": 61200 } ], "logging_steps": 30, "max_steps": 130682, "num_input_tokens_seen": 6266387740, "num_train_epochs": 80, "save_steps": 400, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }