List your company

Judge Accuracy on RMB-SafeRLHF (Target)

69AccuracySeed Rubric
63.90465.22766.5567.873Feb 14, 2026

Evaluation results

MethodLinks
Seed RubricJudge Model=Qwen3-14B2026.02690.127
Biased Rubric SearchJudge Model=Qwen3-14B2026.0267.40.182
Seed RubricJudge Model=Qwen3-14B2026.0267.30.022
Biased Rubric SearchJudge Model=Qwen3-14B2026.0264.10.062