Chinese Legal Reasoning on PLawBench
62.82LLM-as-a-judge ScoreQUBRIC
Evaluation Results
| Method | Links | |
|---|---|---|
| QUBRIC2026.06 | 62.82 | |
| Rubric-RLQuery=Original Query2026.06 | 56.41 | |
| Math-RLVRLength Penalty=0.52026.06 | 54.01 | |
| SFT Baseline2026.06 | 53.97 | |
| AutoIF-RLVRTraining Domain=IF-domain2026.06 | 52.56 | |
| Math-RLVRLength Penalty=none2026.06 | 50.96 |