Reasoning on MATH X-Domain (ID)
83.4AccuracyGPT-4.5
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4.52026.06 | 83.4 | |
| IGA (Ours)2026.06 | 76.4 | |
| Pareto-GS2026.06 | 71.9 | |
| SAND-Mask2026.06 | 71 | |
| PCGrad2026.06 | 70.3 | |
| V-REx2026.06 | 69.8 | |
| IRM2026.06 | 69.3 | |
| ERM-SFT2026.06 | 67.8 | |
| CoT-Distill2026.06 | 67.1 | |
| LoRA-SFT2026.06 | 66.2 |