Multi-hop Question Answering on MuSiQue Wiki-18
24.4Exact Match (EM)Esearch-base
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Esearch-baseModel Backbone=Qwen2.5-7b, Search Engine=Wiki-18, Model Variant=Base2025.10 | 24.4 | 37.1 | |
| Esearch-instructModel Backbone=Qwen2.5-7b, Search Engine=Wiki-18, Model Variant=Instruct2025.10 | 24.1 | 35.8 | |
| Esearch-baseModel Backbone=Qwen2.5-3b, Search Engine=Wiki-18, Model Variant=Base2025.10 | 23.6 | 34.5 | |
| Esearch-instructModel Backbone=Qwen2.5-3b, Search Engine=Wiki-18, Model Variant=Instruct2025.10 | 23.2 | 33.9 | |
| RAGModel Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 8.9 | 14.8 | |
| R1-instructModel Backbone=Qwen2.5-3b, Search Engine=Wiki-18, Model Variant=Instruct2025.10 | 8.5 | 15.1 | |
| RAGModel Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 8.4 | 14.4 | |
| SFTModel Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 8 | 13.8 | |
| SFTModel Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 7.7 | 13.9 | |
| R1-baseModel Backbone=Qwen2.5-3b, Search Engine=Wiki-18, Model Variant=Base2025.10 | 7 | 12.7 | |
| R1-baseModel Backbone=Qwen2.5-7b, Search Engine=Wiki-18, Model Variant=Base2025.10 | 6.6 | 15.7 | |
| R1-instructModel Backbone=Qwen2.5-7b, Search Engine=Wiki-18, Model Variant=Instruct2025.10 | 6 | 14.6 | |
| IRCoTModel Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 5.8 | 14.1 | |
| IRCoTModel Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 5.5 | 14.3 | |
| Search-o1Model Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 4.7 | 12.6 | |
| Search-o1Model Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 4.5 | 13.2 | |
| CoTModel Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 3.5 | 9.4 | |
| Direct InferenceModel Backbone=Qwen2.5-7b, Search Engine=Wiki-182025.10 | 1.9 | 10.6 | |
| Direct InferenceModel Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 1.4 | 9.5 | |
| CoTModel Backbone=Qwen2.5-3b, Search Engine=Wiki-182025.10 | 0.9 | 6.7 |