List your company

Multi-Hop Question Answering on Musique 2018 Wikipedia dump (dev)

22.1AccuracyMR-Search
1.1966.62312.0517.477Mar 11, 2026

Evaluation results

MethodLinks
MR-SearchBackbone=Qwen2.5-7b2026.0322.1
StepResearchBackbone=Qwen2.5-7b2026.0320.5
Search-R1Backbone=Qwen2.5-7b2026.0318.1
ReSearchBackbone=Qwen2.5-7b2026.0316.6
MR-SearchBackbone=Qwen2.5-3b2026.0316.5
PPRMBackbone=Qwen2.5-7b2026.0314.7
PPRMBackbone=Qwen2.5-3b2026.0312.7
StepResearchBackbone=Qwen2.5-3b2026.0310.5
Search-R1Backbone=Qwen2.5-3b2026.037.7
ReSearchBackbone=Qwen2.5-3b2026.037.4
Search-o1Backbone=Qwen2.5-7b2026.035.8
Search-o1Backbone=Qwen2.5-3b2026.035.4
Direct InferenceBackbone=Qwen2.5-7b2026.033.1
Direct InferenceBackbone=Qwen2.5-3b2026.032