Mathematical Reasoning on MetaMathQA 100k-samples
0.142LossLoRA
Evaluation Results
| Method | Links | |
|---|---|---|
| LoRABackbone model=Qwen-2.5-3B, Fine-tuning protocol=fine-tuning2026.05 | 0.142 | |
| DORABackbone model=Qwen-2.5-3B, Fine-tuning protocol=fine-tuning2026.05 | 0.142 | |
| LoRA-GABackbone model=Qwen-2.5-3B, Fine-tuning protocol=fine-tuning2026.05 | 0.143 | |
| BALORABackbone model=Qwen-2.5-3B, Fine-tuning protocol=fine-tuning2026.05 | 0.144 | |
| OLORABackbone model=Qwen-2.5-3B, Fine-tuning protocol=fine-tuning2026.05 | 0.144 |