Long-context retrieval and reasoning on RULER 8k context
100Score 1 (S1)
Transformer
Updated 24d ago
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
2026.03 | 100 | 100 | 95 | 87 | 37 | 41 | 44 | 0.6 | 37 | 28 | 27 | |
2026.03 | 100 | 66 | 24 | 31 | 0.4 | 24 | 22 | 6.7 | 11 | 20 | 12 | |
2026.03 | 100 | 100 | 86 | 86 | 12 | 62 | 57 | 0.2 | 12 | 31 | 23 | |
2026.03 | 100 | 94 | 12 | 28 | 0.4 | 22 | 14 | 37 | 28 | 22 | 16 | |
2026.03 | 100 | 90 | 4.6 | 38 | 0.2 | 22 | 26 | 0.5 | 32 | 25 | 16 | |
2026.03 | 100 | 100 | 60 | 90 | 45 | 64 | 74 | 26 | 1.1 | 26 | 28 | |
2026.03 | 100 | 100 | 37 | 79 | 18 | 33 | 29 | 0.9 | 2.6 | 22 | 26 |