Direct-Answer Question Answering on RefMem-Bench
32.9Accuracy
REMIND
Updated 1d ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.05 | 32.9 | 55.4 | 27.6 | 30.4 | |
2026.05 | 23.6 | 40.4 | 19.5 | 22.4 | |
2026.05 | 23.2 | 39.9 | 20.1 | 23.2 | |
2026.05 | 23.2 | 36.8 | 19.9 | 23 | |
2026.05 | 22.9 | 39.6 | 16.4 | 18.5 | |
2026.05 | 21.8 | - | 17.5 | 19.9 | |
2026.05 | 21.1 | 37.9 | 17.8 | 21.3 | |
2026.05 | 21.1 | - | 19.9 | 22.3 | |
2026.05 | 19.8 | 38.7 | 17 | 19.8 | |
2026.05 | 19.1 | 39.1 | 15.4 | 18.8 | |
2026.05 | 19 | 35.6 | 18.6 | 21.2 | |
2026.05 | 17.3 | 34.3 | 16.5 | 19.3 | |
2026.05 | 12.7 | 24.6 | 10.7 | 12.9 | |
2026.05 | 12.6 | 29.7 | 12.1 | 14.2 |