Mathematical Reasoning on AIME 25 (Avg@8, Pass@8, Response Length)
2.92Avg@8PowerOPD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PowerOPDalpha=500, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 2.92 | 13.33 | 2,819 | |
| PowerOPDalpha=1, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 2.5 | 10 | 3,703 | |
| Full-vocabTeacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 1.67 | 3.33 | 4,096 | |
| VanillaParam=+ Clip, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 1.25 | 6.67 | 3,842 | |
| PowerOPDalpha=0.5, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 1.25 | 6.67 | 3,662 | |
| PowerOPDalpha=10, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 1.25 | 3.33 | 3,612 | |
| VanillaTeacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.83 | 3.33 | 4,085 | |
| PowerOPDalpha=5, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.83 | 6.67 | 3,629 | |
| PowerOPDalpha=50, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.83 | 6.67 | 3,502 | |
| PowerOPDalpha=100, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.83 | 6.67 | 3,324 | |
| VanillaParam=+ Tanh, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.42 | 3.33 | 3,880 | |
| VanillaParam=+ Z-Score, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.42 | 3.33 | 3,703 | |
| PowerOPDalpha=0.1, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 0.42 | 3.33 | 3,806 |