Mathematical Reasoning on AIME24 (Avg@8, Pass@8, Response Length)
10.83Avg@8Full-vocab KL OPD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Full-vocab KL OPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base2026.06 | 10.83 | 23.33 | 4,096 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=502026.06 | 10.83 | 23.33 | 3,227 | |
| Vanilla OPD + TanhTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base2026.06 | 10.42 | 26.67 | 3,318 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=12026.06 | 10.42 | 23.33 | 3,368 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=5002026.06 | 9.58 | 26.67 | 2,776 | |
| Vanilla OPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base2026.06 | 9.17 | 16.67 | 3,388 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=0.12026.06 | 8.75 | 20 | 3,351 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=52026.06 | 8.33 | 20 | 3,366 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=102026.06 | 8.33 | 26.67 | 3,292 | |
| PowerOPDTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base, alpha=1002026.06 | 8.33 | 20 | 3,011 | |
| Vanilla OPD + ClipTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base2026.06 | 4.17 | 23.33 | 3,342 | |
| Vanilla OPD + Z-ScoreTeacher=Qwen3-4B, Student=Qwen3-1.7B-Base2026.06 | 4.17 | 13.33 | 3,466 |