Multi-Hop QA on 2Wiki out-of-domain (test)
37AccuracyGiGPO
Evaluation results
| Method | Links | |
|---|---|---|
| GiGPO2026.06 | 37 | |
| ECPO2026.06 | 36.6 | |
| StepSearch2026.06 | 32 | |
| Search-R12026.06 | 31.9 | |
| ZeroSearch2026.06 | 30 | |
| R1-Instruct2026.06 | 29.2 |
| Method | Links | |
|---|---|---|
| GiGPO2026.06 | 37 | |
| ECPO2026.06 | 36.6 | |
| StepSearch2026.06 | 32 | |
| Search-R12026.06 | 31.9 | |
| ZeroSearch2026.06 | 30 | |
| R1-Instruct2026.06 | 29.2 |