Long-context language modeling evaluation on RULER
100Single-key Accuracy
Full Attention
Updated 1mo ago
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
2026.03 | 100 | 100 | 100 | 100 | 82.5 | 84.4 | 89.8 | 59.2 | 89.5 | |
2026.03 | 100 | 100 | 100 | 99.5 | 81.2 | 80.3 | 89.8 | 59.2 | 88.8 | |
2026.03 | 100 | 100 | 96.4 | 100 | 78 | 83.7 | 81.6 | 59.2 | 87.4 | |
2026.03 | 95.9 | 95.9 | 97.5 | 97.5 | 76.3 | 76.9 | 79.6 | 59.2 | 84.8 | |
2026.03 | 95.9 | 98 | 92.9 | 93.9 | 85.7 | 76.2 | 75.5 | 59.2 | 84.7 | |
2026.03 | 93.9 | 100 | 94.9 | 94.9 | 73.1 | 89.8 | 77.6 | 51 | 84.4 | |
2026.03 | 91.8 | 100 | 100 | 99 | 79.6 | 87.1 | 81.6 | 59.2 | 87.3 | |
2026.03 | 63.3 | 100 | 98.5 | 98.5 | 82 | 93.9 | 79.6 | 51 | 83.3 |