Question Answering with Abstention on Unified Abstention Benchmark Suite
91.5MMLU Accuracy
TRACE INVERSION
Updated 2mo ago
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2026.04 | 91.5 | 88.3 | 81.2 | 87 | 61.6 | 73.1 | 61.2 | 65.3 | 71.3 | 71.2 | 60.2 | 67.6 | 73.3 | |
2026.04 | 88.5 | 85.1 | 77.8 | 83.8 | 58.5 | 81.4 | 64 | 71.2 | 71.1 | 80.4 | 69.5 | 73.7 | 76.2 | |
2026.04 | 84.9 | 71.5 | 72.2 | 76.2 | 70.7 | 71.8 | 29.8 | 57.3 | 48.8 | 51.1 | 42 | 47.3 | 60.4 | |
2026.04 | 78.8 | 84.7 | 61.4 | 75 | 61.6 | 69.1 | 59 | 63.2 | 58.8 | 61.1 | 48.7 | 56.2 | 64.8 | |
2026.04 | 78.4 | 64.7 | 55.6 | 66.2 | 32.9 | 48.8 | 30.1 | 37.3 | 58.3 | 60.5 | 33.8 | 50.9 | 51.5 | |
2026.04 | 77 | 73.9 | 60 | 70.3 | 65.3 | 62.2 | 41.2 | 56.2 | 50.3 | 64.4 | 48.7 | 54.5 | 60.3 | |
2026.04 | 76.7 | 79.9 | 57.7 | 71.4 | 61.7 | 63.2 | 41.7 | 55.5 | 61.4 | 62.4 | 40.8 | 54.9 | 60.6 | |
2026.04 | 76.5 | 78.4 | 60.1 | 71.7 | 55.7 | 45.4 | 31.7 | 44.3 | 51.1 | 67.2 | 40.4 | 52.9 | 56.3 | |
2026.04 | 75.9 | 73.9 | 53.3 | 67.7 | 43.8 | 62.3 | 30.5 | 45.5 | 61.5 | 64.4 | 41.3 | 55.7 | 56.3 | |
2026.04 | 74.9 | 81.7 | 52 | 69.5 | 70.4 | 81.2 | 31.2 | 60.9 | 60.7 | 59 | 38.8 | 52.8 | 61.1 | |
2026.04 | 74.8 | 74.4 | 63.9 | 71 | 63.3 | 61.1 | 51 | 58.5 | 51 | 61.8 | 31 | 47.9 | 59.1 | |
2026.04 | 74.1 | 71.1 | 51.2 | 65.5 | 32.9 | 55.1 | 39.7 | 42.6 | 45.6 | 64.1 | 49.8 | 53.2 | 53.8 | |
2026.04 | 72.5 | 80.4 | 50.1 | 67.7 | 55.2 | 68.2 | 41.4 | 54.9 | 58.8 | 59 | 42.6 | 53.5 | 58.7 | |
2026.04 | 71.9 | 85 | 78.8 | 78.6 | 78.9 | 71.2 | 71.7 | 73.9 | 67 | 73.4 | 66.8 | 69.1 | 73.8 | |
2026.04 | 71.2 | 73.3 | 75.7 | 73.4 | 66.3 | 83 | 69.4 | 72.9 | 64.9 | 71 | 61.4 | 64.3 | 70.2 | |
2026.04 | 71.1 | 68.4 | 60.1 | 66.5 | 47.3 | 51.3 | 31.9 | 43.5 | 45.1 | 64.7 | 40.1 | 50 | 53.3 | |
2026.04 | 68.9 | 73.1 | 63.9 | 68.6 | 61.8 | 61 | 30.5 | 51.1 | 41.5 | 62.1 | 41.1 | 48.2 | 56 | |
2026.04 | 68.8 | 74.7 | 65.6 | 69.7 | 50.9 | 67.2 | 48.8 | 55.6 | 66.7 | 70.4 | 49 | 62 | 62.4 | |
2026.04 | 63.7 | 72.5 | 42 | 59.4 | 51.6 | 61.8 | 34.5 | 49.3 | 32.2 | 60.3 | 42.4 | 45 | 51.2 | |
2026.04 | 57.8 | 54.7 | 51.6 | 54.7 | 42.6 | 69 | 53.3 | 55 | 46.7 | 60 | 31.2 | 46 | 51.9 | |
2026.04 | 47.7 | 50.9 | 48.8 | 49.1 | 45.1 | 66.6 | 30.3 | 47.3 | 51.2 | 62.4 | 33.2 | 48.9 | 48.4 | |
2026.04 | 47.1 | 50.4 | 50.6 | 49.4 | 55 | 61.2 | 30.7 | 49 | 55.2 | 61.8 | 29.9 | 49 | 49.1 | |
2026.04 | 42.4 | 68.5 | 42 | 51 | 50.4 | 71.8 | 41.4 | 54.5 | 36.9 | 59.8 | 42.2 | 46.3 | 50.6 | |
2026.04 | 37.9 | 54.1 | 43.8 | 45.3 | 63.3 | 77.1 | 40.5 | 60.3 | 45.5 | 51.5 | 41.1 | 46 | 50.6 |