Reasoning on Comp-Math
76.16AccuracyQwen3-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-4BLoss function=base+aux2026.02 | 76.16 | |
| Qwen3-4BLoss function=base2026.02 | 75.2 | |
| Qwen3-8BLoss function=base+aux2026.02 | 75.09 | |
| Qwen3-8BLoss function=base2026.02 | 73.65 | |
| LLaMA3.1-8BLoss function=base+aux2026.02 | 58.52 | |
| LLaMA3.1-8BLoss function=base2026.02 | 57.03 |