List your company

Question Answering on PopQA (EM, F1, SD, SE)

46.3Exact Match (EM)AutoSearch
5.01215.73126.4537.169Apr 19, 2026

Evaluation results

MethodLinks
AutoSearchBackbone=Qwen2.5-7B2026.0446.350.31.0940.8
AutoSearchBackbone=Qwen2.5-3B2026.044650.11.0742.9
Search-R1-baseBackbone=Qwen2.5-7B2026.0445.749.81.9523.4
Search-R1-instructBackbone=Qwen2.5-7B2026.0440.545.41.7922.6
StepSearch-baseBackbone=Qwen2.5-7B2026.0438.6441.5724.7
Search-R1-baseBackbone=Qwen2.5-3B2026.043741.31.3527.4
Search-R1-instructBackbone=Qwen2.5-3B2026.0435.741.21.7220.7
StepSearch-instructBackbone=Qwen2.5-3B2026.0434.940.21.4524.1
StepSearch-instructBackbone=Qwen2.5-7B2026.0434.940.11.9817.6
HIPRAG-instructBackbone=Qwen2.5-7B2026.0434.242.41.1629.5
StepSearch-baseBackbone=Qwen2.5-3B2026.0432.237.12.9410.9
HIPRAG-baseBackbone=Qwen2.5-3B2026.0414.9231.688.88
HIPRAG-instructBackbone=Qwen2.5-3B2026.046.613.71.354.88