List your company

Multi-Hop Question Answering on MuSiQue (out-of-domain)

22.6AccuracyStepSearch
4.0888.89413.718.506Feb 9, 2026

Evaluation results

MethodLinks
StepSearchBackbone=Qwen2.5-7B-Instruct2026.0222.6
SKILLRLBackbone=Qwen2.5-7B-Instruct2026.0220.2
ZeroSearchBackbone=Qwen2.5-7B-Instruct2026.0218.4
EvolveRBackbone=Qwen2.5-7B-Instruct2026.0215.6
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0214.6
RAGBackbone=Qwen2.5-7B-Instruct2026.029.4
Search-o1Backbone=Qwen2.5-7B-Instruct2026.028.6
CoTBackbone=Qwen2.5-7B-Instruct2026.026.6
R1-InstructBackbone=Qwen2.5-7B-Instruct2026.026
Qwen2.5Backbone=Qwen2.5-7B-Instruct2026.024.8