Mathematical Reasoning on AIME (Error Attribution Scores)
46Random ScoreQWEN3 8B
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| QWEN3 8Bn=82026.04 | 46 | 46.9 | 46.1 | 45.5 | 49.5 | 50.3 | 48.9 | 43.7 | |
| QWEN3 8Bn=322026.04 | 45.8 | 47.7 | 46.9 | 39 | 51.2 | 54.9 | 50.4 | 38.3 | |
| QWEN3 8Bn=42026.04 | 45.7 | 46.6 | 45.9 | 45.9 | 48.8 | 48.8 | 48.2 | 44.4 | |
| QWEN3 8Bn=162026.04 | 45.3 | 47.1 | 46.1 | 41.7 | 50.4 | 51.6 | 49.2 | 44.1 | |
| QWEN3 4Bn=162026.04 | 40.4 | 42.4 | 41.1 | 37.8 | 42.3 | 48.3 | 45.7 | 33.3 | |
| QWEN3 4Bn=42026.04 | 40.2 | 41.1 | 40.4 | 41.1 | 40.7 | 43.1 | 41 | 40 | |
| QWEN3 4Bn=322026.04 | 39.7 | 48 | 42.2 | 36.4 | 48.9 | 50.5 | 52.3 | 31.9 | |
| QWEN3 4Bn=82026.04 | 39 | 41.8 | 40.6 | 39.1 | 42.5 | 45.5 | 40.3 | 36.7 | |
| DISTILL-QWEN 7Bn=42026.04 | 24.7 | 27.2 | 25.4 | 21.2 | 32.4 | 31.9 | 26.7 | 24.5 | |
| DISTILL-QWEN 7Bn=322026.04 | 24.4 | 29.7 | 29.7 | 18.6 | 35.1 | 34.2 | 31.1 | 20.2 | |
| DISTILL-QWEN 7Bn=82026.04 | 24.3 | 27.7 | 25.7 | 20.4 | 32.7 | 32.5 | 29.2 | 24 | |
| DISTILL-QWEN 7Bn=162026.04 | 24.1 | 28.1 | 26.6 | 20.1 | 33.1 | 33.1 | 29.5 | 22.7 | |
| DISTILL-QWEN 1.5Bn=82026.04 | 18.4 | 18.1 | 19.6 | 12.4 | 22.1 | 22.2 | 20.8 | 18.3 | |
| DISTILL-QWEN 1.5Bn=322026.04 | 18.3 | 20.4 | 20 | 13.3 | 24.6 | 22.8 | 22.1 | 17.5 | |
| DISTILL-QWEN 1.5Bn=42026.04 | 18.1 | 18.4 | 19.2 | 13.6 | 21 | 21.8 | 20.2 | 18.1 | |
| DISTILL-QWEN 1.5Bn=162026.04 | 18.1 | 18.8 | 19.9 | 12.2 | 22.3 | 22.4 | 21.1 | 17.6 |