Question Answering on TriviaQA (test) (EM, F1)
60EMR2-Searcher
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| R2-Searcher2026.06 | 60 | 67.5 | |
| AutoRefine2026.06 | 58.1 | 66.9 | |
| ReSearch2026.06 | 57.4 | 65.5 | |
| ZeroSearch2026.06 | 52.8 | 60.3 | |
| Search-R12026.06 | 52.5 | 59.7 | |
| Search-o12026.06 | 42.8 | 50.1 | |
| RAG2026.06 | 42.4 | 47.8 | |
| IRCOT2026.06 | 38.4 | 48.2 | |
| COT2026.06 | 31.2 | 36.3 | |
| Direct Inference2026.06 | 29.4 | 34.3 |