Multi-round Co-reference Resolution on Long-context benchmarks
38.5Score (8k Context)
CLP
Updated 1mo ago
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
2026.03 | 38.5 | 33.9 | 31.9 | 33.9 | 30.1 | |
2026.03 | 36.3 | 34.1 | 31.2 | 34.6 | 26.7 | |
2026.03 | 35.7 | 35.5 | 33.4 | 29.2 | 17.2 | |
2026.03 | 35.7 | 34.9 | 32.5 | 29.9 | 8.9 | |
2026.03 | 35 | 26.1 | 16.6 | 13.7 | 8.7 | |
2026.03 | 34.6 | 30.4 | 28.9 | 26.3 | 21.3 | |
2026.03 | 34 | 29.8 | 31.9 | 31.6 | 25.5 | |
2026.03 | 32.9 | 31.5 | 25.8 | 24.2 | 19.3 | |
2026.03 | 32.5 | 31.8 | 26.9 | 25.9 | 19 | |
2026.03 | 32.4 | 30.7 | 30.8 | 21 | 11.4 | |
2026.03 | 32.4 | 31.5 | 29.6 | 25.8 | 11.6 | |
2026.03 | 32 | 30.9 | 29.9 | 22.4 | 14.8 | |
2026.03 | 29.7 | 27 | 28.3 | 24.5 | 17 | |
2026.03 | 26.8 | 22.2 | 19 | 17.8 | 12.8 | |
2026.03 | 26.7 | 22.8 | 17.6 | 13.9 | 11 | |
2026.03 | 26.4 | 20.7 | 16 | 15.6 | 9.4 | |
2026.03 | 26.4 | 22.7 | 18 | 15.4 | 5.6 | |
2026.03 | 26.3 | 22.8 | 19.7 | 19.1 | 14.3 | |
2026.03 | 26.3 | 23.1 | 18.6 | 18.7 | 8.9 | |
2026.03 | 25.6 | 18.7 | 8.3 | 9.7 | 6.1 | |
2026.03 | 24 | 19.4 | 16.5 | 13 | 12.1 |