Multi-Hop Question Answering on HotpotQA in-domain (val test)
49.9Exact Match (EM)Search-R2
Evaluation Results
| Method | Links | |
|---|---|---|
| Search-R2Backbone=Qwen2.5-32B2026.02 | 49.9 | |
| Search-R1Backbone=Qwen2.5-32B2026.02 | 43.3 | |
| Search-R2Backbone=Qwen3-8B2026.02 | 41.2 | |
| Search-R2Backbone=Qwen2.5-7B2026.02 | 39 | |
| Search-R1Backbone=Qwen3-8B2026.02 | 37.2 | |
| Rejection SamplingBackbone=Qwen2.5-7B2026.02 | 33.1 | |
| Search-R1Backbone=Qwen2.5-7B2026.02 | 32.6 | |
| RAGBackbone=Qwen2.5-7B2026.02 | 29.9 | |
| R1-baseBackbone=Qwen2.5-7B2026.02 | 24.2 | |
| R1-instructBackbone=Qwen2.5-7B2026.02 | 23.7 | |
| SFTBackbone=Qwen2.5-7B2026.02 | 21.7 | |
| Search-o1Backbone=Qwen2.5-7B2026.02 | 18.7 | |
| Direct InferenceBackbone=Qwen2.5-7B2026.02 | 18.3 | |
| IRCoTBackbone=Qwen2.5-7B2026.02 | 13.3 | |
| CoTBackbone=Qwen2.5-7B2026.02 | 9.2 |