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 / methodology

5.91 kB

1 contributor

History: 3 commits

CinderD's picture

Fix evaluation harness and normalize rubrics

b37dd4f verified 4 days ago

EVAL_PROTOCOL.md

5.91 kB
Fix evaluation harness and normalize rubrics 4 days ago