List your company

Multi-hop Question Answering on MuSiQue (EM, LJ)

24.12EMM2R
1.767.56513.3719.175Apr 10, 2026

Evaluation results

MethodLinks
M2RBackbone=Qwen2.5-7B-Instruct2026.0424.1235.44
ReSearchBackbone=Qwen2.5-7B-Instruct2026.0422.333.43
M2RBackbone=Qwen2.5-3B-Instruct2026.0420.8732.97
SUREBackbone=Qwen2.5-7B-Instruct2026.0420.8732.24
ReSearchBackbone=Qwen2.5-3B-Instruct2026.0419.431.56
SUREBackbone=Qwen2.5-3B-Instruct2026.0417.2427.61
COFTBackbone=Qwen2.5-7B-Instruct2026.0417.1226.28
COFTBackbone=Qwen2.5-3B-Instruct2026.0413.4920.11
Iter-RetGenBackbone=Qwen2.5-7B-Instruct2026.048.6916.14
IRCoTBackbone=Qwen2.5-7B-Instruct2026.046.9914.19
Iter-RetGenBackbone=Qwen2.5-3B-Instruct2026.046.9113.43
Naive RAGBackbone=Qwen2.5-7B-Instruct2026.046.2112.78
IRCoTBackbone=Qwen2.5-3B-Instruct2026.044.3910.83
Naive RAGBackbone=Qwen2.5-3B-Instruct2026.044.1410.32
Naive GenerationBackbone=Qwen2.5-7B-Instruct2026.043.7610.38
Naive GenerationBackbone=Qwen2.5-3B-Instruct2026.042.626.08