List your company

Multi-Hop Question Answering on 2WikiMultiHopQA out-of-domain (val test)

51.7Exact Match (EM)Search-R2
9.47620.43831.442.362Feb 3, 2026

Evaluation results

MethodLinks
Search-R2Backbone=Qwen2.5-32B2026.0251.7
Search-R1Backbone=Qwen2.5-32B2026.0246.2
Search-R2Backbone=Qwen3-8B2026.0240.5
Search-R2Backbone=Qwen2.5-7B2026.0235.8
Search-R1Backbone=Qwen3-8B2026.0235.5
Search-R1Backbone=Qwen2.5-7B2026.0229.7
Rejection SamplingBackbone=Qwen2.5-7B2026.0229.6
R1-instructBackbone=Qwen2.5-7B2026.0229.2
R1-baseBackbone=Qwen2.5-7B2026.0227.3
SFTBackbone=Qwen2.5-7B2026.0225.9
Direct InferenceBackbone=Qwen2.5-7B2026.0225
RAGBackbone=Qwen2.5-7B2026.0223.5
Search-o1Backbone=Qwen2.5-7B2026.0217.6
IRCoTBackbone=Qwen2.5-7B2026.0214.9
CoTBackbone=Qwen2.5-7B2026.0211.1