Multi-Hop QA on Bamboogle (out-of-domain, test)
34.4AccuracyStepSearch
Evaluation results
| Method | Links | |
|---|---|---|
| StepSearch2026.06 | 34.4 | |
| R1-Instruct2026.06 | 29.3 | |
| ECPO2026.06 | 27.2 | |
| Search-R12026.06 | 26.4 | |
| GiGPO2026.06 | 26.4 | |
| ZeroSearch2026.06 | 11.1 |
| Method | Links | |
|---|---|---|
| StepSearch2026.06 | 34.4 | |
| R1-Instruct2026.06 | 29.3 | |
| ECPO2026.06 | 27.2 | |
| Search-R12026.06 | 26.4 | |
| GiGPO2026.06 | 26.4 | |
| ZeroSearch2026.06 | 11.1 |