Judge Accuracy on Ultra-Problem (Bench)
73AccuracyBiased Rubric Search
Evaluation Results
| Method | Links | |
|---|---|---|
| Biased Rubric SearchJudge Model=Qwen3-14B2026.02 | 73 | |
| Seed RubricJudge Model=Qwen3-14B2026.02 | 72.8 |
| Method | Links | |
|---|---|---|
| Biased Rubric SearchJudge Model=Qwen3-14B2026.02 | 73 | |
| Seed RubricJudge Model=Qwen3-14B2026.02 | 72.8 |