Single-Hop Question Answering on NQ (EM, F1, CEM)
47.1EMMultiSearch
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MultiSearchBackbone=3B-Base2026.05 | 47.1 | 56.4 | 51.4 | |
| AutoRefine2026.05 | 46.7 | 53.4 | 50.2 | |
| Search-R12026.05 | 42.1 | 47.6 | 46.2 |
Loading the SOTA2 catalog…
| Method | Links | |||
|---|---|---|---|---|
| MultiSearchBackbone=3B-Base2026.05 | 47.1 | 56.4 | 51.4 | |
| AutoRefine2026.05 | 46.7 | 53.4 | 50.2 | |
| Search-R12026.05 | 42.1 | 47.6 | 46.2 |