Question Answering on MuSiQue (EM, F1, Trigger rate (T))
21.8Exact Match (EM)Verbal-Calibrate
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Verbal-CalibrateScope=global2026.04 | 21.8 | 28.8 | 76.8 | |
| Uncertain-CalibrateScope=local2026.04 | 17.6 | 24.1 | 94.2 | |
| DRAGIN2026.04 | 17 | 27.1 | 87 | |
| Base-Verbal2026.04 | 10.8 | 17.5 | 13.6 | |
| Ret-All2026.04 | 9.8 | 17.4 | 100 | |
| ADARAGUE2026.04 | 9.6 | 15.9 | 53.8 | |
| Base-UncTok2026.04 | 8.8 | 16.2 | 4.3 | |
| FLARE2026.04 | 6.4 | 13.8 | 99.8 | |
| No-Ret2026.04 | 5.6 | 10.3 | — | |
| SR-13BBackbone=13B2026.04 | 0.8 | 6.1 | 39.8 | |
| SR-7BBackbone=7B2026.04 | 0.6 | 5.2 | 56 |