List your company

Monolingual Question Answering on POPQA

65.04fEMCroSearch-R1
7.798422.659237.5252.3808Apr 28, 2026

Evaluation results

MethodLinks
CroSearch-R1Backbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.0465.0483.46
SFTBackbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.0462.668.4
CroSearch-R1Backbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.0461.3378.04
Search-R1Backbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.0459.9678.39
SFTBackbone=Qwen2.5-3B-Instruct, Paradigm=Post-training2026.0459.565.18
Search-R1Backbone=Qwen2.5-7B-Instruct, Paradigm=Post-training2026.0458.270.96
Search-o1Backbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.0454.0462.5
IRCoTBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.0452.6860.44
RAGBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.0450.159.47
IRCoTBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.0431.4350.22
Search-o1Backbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.0430.4847.06
Direct InferenceBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.0424.937.06
Direct InferenceBackbone=Qwen2.5-3B-Instruct, Paradigm=Training-free2026.0418.530.5
RAGBackbone=Qwen2.5-7B-Instruct, Paradigm=Training-free2026.041055