Mathematical Reasoning on Minerva (Avg@8, Pass@8, response length)
12.22Average Score @8PowerOPD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PowerOPDalpha=5, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 12.22 | 23.63 | 1,129 | |
| PowerOPDalpha=500, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 11.4 | 23.53 | 1,273 | |
| PowerOPDalpha=10, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 11.31 | 23.9 | 1,124 | |
| PowerOPDalpha=50, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 11.31 | 22.43 | 1,149 | |
| Full-vocabTeacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 11.03 | 15.81 | 4,077 | |
| PowerOPDalpha=100, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 10.75 | 23.69 | 1,225 | |
| PowerOPDalpha=0.5, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 10.62 | 23.16 | 1,225 | |
| PowerOPDalpha=1, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 10.57 | 24.26 | 1,236 | |
| VanillaParam=+ Clip, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 10.06 | 21.69 | 1,378 | |
| VanillaParam=+ Tanh, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 10.03 | 23 | 1,329 | |
| PowerOPDalpha=0.1, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 9.38 | 20.96 | 1,391 | |
| VanillaParam=+ Z-Score, Teacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 8 | 20.22 | 1,410 | |
| VanillaTeacher=Qwen3-4B, Student=Qwen3-0.6B-Base2026.06 | 6.99 | 17.65 | 4,065 |