XHotpotQA · GLM-5.2 Judge
AUDIT SNAPSHOT · V1
Sanitized, language-balanced translation-quality audit
2,760 scored units · stable hash joins · public reasoning removed
AUDIT · V1
ROWS · 2,760
SCORED · 100%
MODEL · requested alias only
COMPARISON · independent / unpaired
Navigate:
Overview ·
Load ·
Methodology ·
Results ·
Schema ·
Provenance ·
Limitations ·
Citation ·
Collection… See the full description on the dataset page: https://huggingface.co/datasets/Iman998/XhotpotQA-GLM52-Judge-V1.