List your company

Multi-Hop Question Answering on Musique out-of-domain (val test)

25.4Exact Match (EM)Search-R2
1.2727.53613.820.064Feb 3, 2026

Evaluation results

MethodLinks
Search-R2Backbone=Qwen2.5-32B2026.0225.4
Search-R1Backbone=Qwen2.5-32B2026.0222.1
Search-R2Backbone=Qwen3-8B2026.0217.2
Search-R1Backbone=Qwen3-8B2026.0215.7
Search-R2Backbone=Qwen2.5-7B2026.0215.1
Search-R1Backbone=Qwen2.5-7B2026.0212.5
Rejection SamplingBackbone=Qwen2.5-7B2026.0212.3
R1-baseBackbone=Qwen2.5-7B2026.028.3
IRCoTBackbone=Qwen2.5-7B2026.027.2
R1-instructBackbone=Qwen2.5-7B2026.027.2
SFTBackbone=Qwen2.5-7B2026.026.6
Search-o1Backbone=Qwen2.5-7B2026.025.8
RAGBackbone=Qwen2.5-7B2026.025.8
Direct InferenceBackbone=Qwen2.5-7B2026.023.1
CoTBackbone=Qwen2.5-7B2026.022.2