Capability Self-Assessment on MMLU-Pro Science
69.1M-F1OLMo2-7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OLMo2-7BTraining Strategy=RLVR2026.05 | 69.1 | 10.95 | 106.5 | |
| Qwen3-1.7BTraining Strategy=RLVR2026.05 | 67.4 | 10.12 | 102.2 | |
| OLMo2-7BTraining Strategy=SFTSelf2026.05 | 66.3 | 7.83 | 88.3 | |
| OLMo2-13BTraining Strategy=RLVR2026.05 | 65.7 | 8.4 | 99.5 | |
| OLMo2-7BTraining Strategy=SFTTeacher2026.05 | 65.4 | 7.08 | 76.9 | |
| Qwen3-4BTraining Strategy=RLVR2026.05 | 64.4 | 8.07 | 101.8 | |
| Qwen3-4BTraining Strategy=SFTTeacher2026.05 | 63.6 | 7.38 | 87.6 | |
| OLMo2-13BTraining Strategy=SFTTeacher2026.05 | 63.6 | 6.99 | 61.1 | |
| Qwen3-8BTraining Strategy=RLVR2026.05 | 62.4 | 6.72 | 101.5 | |
| Qwen3-8BTraining Strategy=SFTTeacher2026.05 | 62.3 | 6.34 | 94 | |
| AverageTraining Strategy=RLVR2026.05 | 62 | 7.07 | 99.8 | |
| Qwen3-1.7BTraining Strategy=SFTSelf2026.05 | 61.1 | 6.39 | 89.6 | |
| AverageTraining Strategy=SFTTeacher2026.05 | 61.1 | 5.61 | 80.8 | |
| Qwen3-14BTraining Strategy=SFTSelf2026.05 | 61 | 5.57 | 92.5 | |
| AverageTraining Strategy=SFTSelf2026.05 | 60.8 | 5.65 | 83.8 | |
| OLMo2-13BTraining Strategy=SFTSelf2026.05 | 60.6 | 5.7 | 76.2 | |
| Qwen3-14BTraining Strategy=SFTTeacher2026.05 | 60.4 | 5.12 | 95.2 | |
| Qwen3-4BTraining Strategy=SFTSelf2026.05 | 59.8 | 5.92 | 86.8 | |
| Qwen3-0.6BTraining Strategy=SFTSelf2026.05 | 58.6 | 3.97 | 61.8 | |
| Qwen3-8BTraining Strategy=SFTSelf2026.05 | 58 | 4.19 | 91.5 | |
| Qwen3-0.6BTraining Strategy=SFTTeacher2026.05 | 57.4 | 3.63 | 69.4 | |
| Qwen3-14BTraining Strategy=RLVR2026.05 | 56.6 | 5.25 | 100.4 | |
| OLMo2-7BTraining Strategy=Vanilla2026.05 | 55.1 | 3.44 | 100 | |
| Qwen3-1.7BTraining Strategy=SFTTeacher2026.05 | 54.7 | 2.75 | 81.2 | |
| OLMo2-13BTraining Strategy=Vanilla2026.05 | 54.4 | 6.13 | 100 | |
| Qwen3-8BTraining Strategy=SFTLabel2026.05 | 50.2 | 10.73 | 49.1 | |
| Qwen3-1.7BTraining Strategy=SFTLabel2026.05 | 49.2 | 5.3 | 37.3 | |
| Qwen3-0.6BTraining Strategy=RLVR2026.05 | 48.7 | 0.03 | 86.8 | |
| Qwen3-1.7BTraining Strategy=Vanilla2026.05 | 46.4 | 3.02 | 100 | |
| AverageTraining Strategy=Vanilla2026.05 | 45.4 | 3.06 | 100 | |
| Qwen3-8BTraining Strategy=Vanilla2026.05 | 45 | — | 100 | |
| Qwen3-14BTraining Strategy=Vanilla2026.05 | 45 | — | 100 | |
| Qwen3-0.6BTraining Strategy=SFTLabel2026.05 | 44.9 | 0 | 2 | |
| OLMo2-7BTraining Strategy=SFTLabel2026.05 | 44.4 | 0 | 0.7 | |
| Qwen3-4BTraining Strategy=Vanilla2026.05 | 42.2 | — | 100 | |
| OLMo2-13BTraining Strategy=SFTLabel2026.05 | 42 | 0 | 5 | |
| AverageTraining Strategy=SFTLabel2026.05 | 40.4 | 3.25 | 17.4 | |
| Qwen3-4BTraining Strategy=SFTLabel2026.05 | 31.5 | 4.47 | 20.8 | |
| Qwen3-0.6BTraining Strategy=Vanilla2026.05 | 29.5 | 0.33 | 100 | |
| Qwen3-14BTraining Strategy=SFTLabel2026.05 | 20.5 | 2.25 | 6.7 |