Open-domain Question Answering on TriviaQA Search-o1 framework (50% held-out test)
69.6AccuracyDefault
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DefaultN=2502026.05 | 69.6 | 344 | — | |
| PROBE&PREFILLN=2502026.05 | 69.2 | 275 | -20 | |
| Best BaselineN=2502026.05 | 68.8 | 289 | -16 |
| Method | Links | |||
|---|---|---|---|---|
| DefaultN=2502026.05 | 69.6 | 344 | — | |
| PROBE&PREFILLN=2502026.05 | 69.2 | 275 | -20 | |
| Best BaselineN=2502026.05 | 68.8 | 289 | -16 |