Multi-Hop Question Answering on Bamboogle (EM, F1, CEM)
0.39EMMultiSearch
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MultiSearchBackbone=3B-Base2026.05 | 0.39 | 0.542 | 0.433 | |
| AutoRefine2026.05 | 0.344 | 0.449 | 0.384 | |
| Search-R12026.05 | 0.128 | 0.184 | 0.128 |
| Method | Links | |||
|---|---|---|---|---|
| MultiSearchBackbone=3B-Base2026.05 | 0.39 | 0.542 | 0.433 | |
| AutoRefine2026.05 | 0.344 | 0.449 | 0.384 | |
| Search-R12026.05 | 0.128 | 0.184 | 0.128 |