Reward Modeling on Overall Performance (test)
80.2OverallCE-RM-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| CE-RM-4BRM Type=Pointwise, Scaling=42026.01 | 80.2 | |
| CE-RM-4BRM Type=Pointwise, Scaling=22026.01 | 79.2 | |
| CE-RM-4BRM Type=Pointwise, Scaling=12026.01 | 77.4 | |
| TIR-Judge-Zero-8BRM Type=Pointwise2026.01 | 73.8 | |
| TIR-Judge-Distill-8BRM Type=Pointwise2026.01 | 73.7 | |
| TIR-Judge-Zero-4BRM Type=Pointwise2026.01 | 71.7 | |
| Gemini-2.5-FlashRM Type=Pointwise2026.01 | 71.2 | |
| TIR-Judge-Distill-4BRM Type=Pointwise2026.01 | 69.7 | |
| CompassJudger1-32BRM Type=Pointwise2026.01 | 56.5 |