List your company

Open-domain Question Answering on MuSiQue (ACC, SC)

26.2Accuracy (ACC)StepSearch
5.29610.72316.1521.577May 28, 2026

Evaluation results

MethodLinks
StepSearchBackbone=Qwen2.5-7B-Instruct2026.0526.22.27
HiPRAGBackbone=Qwen2.5-7B-Instruct2026.0523.82.55
RFTBackbone=Qwen2.5-7B-Instruct2026.0523.12.5
SAASBackbone=Qwen2.5-7B-Instruct2026.0522.61.3
SAASBackbone=Qwen2.5-3B-Instruct2026.0520.91.62
RFTBackbone=Qwen2.5-3B-Instruct2026.05172.76
HiPRAGBackbone=Qwen2.5-3B-Instruct2026.05172.37
StepSearchBackbone=Qwen2.5-3B-Instruct2026.0516.82.1
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0516.81.38
Search-R1Backbone=Qwen2.5-3B-Instruct2026.0515.21.79
Direct InferenceBackbone=Qwen2.5-7B-Instruct2026.058.8—
Direct InferenceBackbone=Qwen2.5-3B-Instruct2026.056.1—