Datasets:

CinderD
/

wildtrace

Tasks:

Question Answering

Text Generation

Modalities:

Formats:

Sub-tasks:

document-question-answering

Languages:

Size:

ArXiv:

Tags:

multi-hop-reasoning

source-internal-reasoning

evidence-withheld

Libraries:

License:

Dataset card Data Studio Files Files and versions

wildtrace / eval

Commit History

Aug 13

09:07
Fix evaluation harness and normalize rubrics⁠

CinderD

Verified signature
b37dd4f

Jul 10

09:15
Add Gemini-native judge compatibility to eval harness

CinderD

Verified signature
9c7ab6a

Jun 30

03:34
Release WildTrace strict481 public package

CinderD

Verified signature
950f6b2