List your company

Factual Reasoning on SampleQA (Accuracy)

3.21AccuracyVRPO
1.67082.07042.472.8696Aug 5, 2025

Evaluation results

MethodLinks
VRPOBase Model=Qwen3-8B2025.083.21
KTAEBase Model=Qwen3-8B2025.083.17
λ-GRPOBase Model=Qwen2.5-7B-Cold Start2025.082.94
BaseBase Model=Qwen3-8B2025.082.89
VRPOBase Model=Qwen2.5-7B-Cold Start2025.082.82
λ-GRPOBase Model=Qwen3-8B2025.082.77
BaseBase Model=Qwen2.5-7B-Cold Start2025.082.36
KTAEBase Model=Qwen2.5-7B-Cold Start2025.081.73