agent-reward-bench — Dataset Trust & Quality Analysis 2026

agent-reward-bench is a dataset by McGill-NLP with Nerq Trust Score 53/100 (D). Verdict: caution. This asset has moderate trust signals. Review before using in production.

agent-reward-bench is a dataset by McGill-NLP with a Nerq Trust Score of 53/100 (Grade D). Verdict: CAUTION. 914 downloads. License: Not specified. Last analyzed February 2026.

53
Trust Score
D
Grade
914
Downloads
4
Stars

About This Dataset

N/A

Dataset Details

AuthorMcGill-NLP
SourceHuggingFace
LicenseNot specified
Tagstask_categories:robotics, language:en, size_categories:1K<n<10K, format:csv, modality:image, modality:text, library:datasets, library:pandas, library:mlcroissant, library:polars
LanguageNot specified
Downloads914

Quality & Trust Analysis

Security ScorePending/100
Activity ScorePending/100
DocumentationPending/100
Popularity0.1/100
EU AI Act ClassPending

Similar Datasets

DatasetDownloadsTrust
huginn-dataset501.6K62
tiny-agents32255
piaf19956
AgentTraj-L18056
AFM-WebAgent-RL-Dataset13555
Yunjue-Agent-Traces11852

Frequently Asked Questions

Is agent-reward-bench safe to use?
agent-reward-bench has a Nerq Trust Score of 53/100 (Grade D). Verdict: caution. Always check the license and data provenance before using in production.
What license is agent-reward-bench?
License: Not specified. Verify the license terms on HuggingFace before commercial use.
How popular is agent-reward-bench?
agent-reward-bench has 914 downloads and 4 stars on HuggingFace, indicating growing community adoption.
Who maintains agent-reward-bench?
agent-reward-bench is maintained by McGill-NLP. Activity score: pending/100.
Check Any Dataset
curl nerq.ai/v1/preflight?target=agent-reward-bench
API docs →

Disclaimer: Nerq trust scores are automated assessments. Not endorsements.

We use cookies for analytics and caching. Privacy Policy