List your company

Single-Hop Question Answering on NQ (Natural Questions) in-domain (test)

31.45EMARPO
-1.2587.233515.72524.2165Feb 12, 2026

Evaluation results

MethodLinks
ARPOModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.0231.451.93
SIGHTModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.0231.451.1
Tree-GRPOModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.0230.861.26
SIGHTModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.0229.31.19
ARPOModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.0228.322.93
Search-R1Model=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.0225.981.57
Search-R1Model=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.02251.69
Tree-GRPOModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.0223.241.47
Search-o1Model=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.0219.530.77
Search-o1Model=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.0216.990.9
GRPOModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/o Retrieval2026.0216.210
GRPOModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/o Retrieval2026.0214.650
Vanilla RAGModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.029.961
ReActModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.028.591.45
COTModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/o Retrieval2026.025.080
Direct ReasoningModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/o Retrieval2026.023.710
Vanilla RAGModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.021.561
ReActModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.020.21.89
Direct ReasoningModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/o Retrieval2026.0200
COTModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/o Retrieval2026.0200