List your company

Question Answering on Search-augmented QA tasks Average

48.3Average AccuracyHCAPO (Ours)
21.36428.35735.3542.343Mar 7, 2026

Evaluation results

MethodLinks
HCAPO (Ours)Type=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0348.3
GiGPOType=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0347.2
HCAPO (Ours)Type=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0344.3
GiGPOType=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0342.1
StepSearchType=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0340
ZeroSearchType=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0339.1
Search-R1Type=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0338.5
StepSearchType=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0334.4
Search-R1Type=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0332.5
ZeroSearchType=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0331.7
R1-InstructType=RL Training, Base Model=Qwen2.5-3B-Instruct2026.0327.1
R1-InstructType=RL Training, Base Model=Qwen2.5-7B-Instruct2026.0322.4