Multi-hop Reasoning on MuSiQue IRCoT 500 samples (test)
25.59ACCActiShade
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ActiShadeBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Active/Iterative retrieval2026.01 | 25.59 | 27.47 | |
| ActiShadeBackbone=Llama-3-8B-Instruct, Retrieval Mode=Active/Iterative retrieval2026.01 | 25.25 | 26.94 | |
| ActiShadeBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Active/Iterative retrieval2026.01 | 22.8 | 26.11 | |
| DRAGINBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 22.7 | 24.11 | |
| DRAGINBackbone=Llama-3-8B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 21.11 | 22.61 | |
| SelfASKBackbone=Llama-3-8B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 20.6 | 21.41 | |
| DRAGINBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 19.8 | 22.01 | |
| FLAREBackbone=Llama-3-8B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 19.74 | 20.5 | |
| SelfASKBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 18.48 | 21.75 | |
| Iter-RetGenBackbone=Llama-3-8B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 18.24 | 20.59 | |
| SelfASKBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 17.35 | 20.6 | |
| Iter-RetGenBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 17.2 | 21.54 | |
| IRCoTBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 15.89 | 19.9 | |
| IRCoTBackbone=Llama-3-8B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 15.57 | 18.32 | |
| Iter-RetGenBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 15.4 | 18.07 | |
| IRCoTBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 14.9 | 18.01 | |
| Direct-RBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Single-round retrieval2026.01 | 14.8 | 19.16 | |
| FLAREBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 13.1 | 18 | |
| CoTBackbone=Llama-3-8B-Instruct, Retrieval Mode=None2026.01 | 11.65 | 16.29 | |
| Direct-RBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Single-round retrieval2026.01 | 11.6 | 17.24 | |
| Direct-RBackbone=Llama-3-8B-Instruct, Retrieval Mode=Single-round retrieval2026.01 | 11.42 | 16.06 | |
| FLAREBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=Multi-round retrieval2026.01 | 10.69 | 14.89 | |
| CoTBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=None2026.01 | 10.4 | 17.74 | |
| DirectBackbone=Qwen2.5-14B-Instruct, Retrieval Mode=None2026.01 | 6.2 | 13.48 | |
| CoTBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=None2026.01 | 6 | 13.93 | |
| DirectBackbone=Llama-3-8B-Instruct, Retrieval Mode=None2026.01 | 5.6 | 9.96 | |
| DirectBackbone=Qwen2.5-7B-Instruct, Retrieval Mode=None2026.01 | 3.8 | 11.09 |