Open-domain Question Answering on NQ Search-o1 framework (50% held-out test)
44.8AccuracyDefault
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DefaultN=2502026.05 | 44.8 | 263 | — | |
| Best BaselineN=2502026.05 | 44.8 | 252 | -4 | |
| PROBE&PREFILLN=2502026.05 | 42 | 248 | -6 |
| Method | Links | |||
|---|---|---|---|---|
| DefaultN=2502026.05 | 44.8 | 263 | — | |
| Best BaselineN=2502026.05 | 44.8 | 252 | -4 | |
| PROBE&PREFILLN=2502026.05 | 42 | 248 | -6 |