General Question Answering on TriviaQA (OSR%)
12.2OSR (%)StepSearch-base
Evaluation Results
| Method | Links | |
|---|---|---|
| StepSearch-baseBackbone=Qwen2.5-3B2026.04 | 12.2 | |
| StepSearch-instructBackbone=Qwen2.5-7B2026.04 | 10.9 | |
| StepSearch-baseBackbone=Qwen2.5-7B2026.04 | 7.1 | |
| Search-R1-baseBackbone=Qwen2.5-7B2026.04 | 6.5 | |
| StepSearch-instructBackbone=Qwen2.5-3B2026.04 | 4.3 | |
| Search-R1-instructBackbone=Qwen2.5-3B2026.04 | 3.9 | |
| Search-R1-instructBackbone=Qwen2.5-7B2026.04 | 3.9 | |
| HIPRAG-instructBackbone=Qwen2.5-7B2026.04 | 2.2 | |
| Search-R1-baseBackbone=Qwen2.5-3B2026.04 | 2.1 | |
| HIPRAG-baseBackbone=Qwen2.5-3B2026.04 | 1.8 | |
| HIPRAG-instructBackbone=Qwen2.5-3B2026.04 | 0.8 | |
| AutoSearchBackbone=Qwen2.5-7B2026.04 | 0.51 | |
| AutoSearchBackbone=Qwen2.5-3B2026.04 | 0.1 |