List your company

Standard Question Answering on SQuAD v2

19.53EMPrompt-R1
4.90768.703812.516.2962Nov 2, 2025

Evaluation results

MethodLinks
Prompt-R12025.1119.5329.28
CoT ReasoningBackbone=GPT-4o-mini2025.1114.0625.73
BaselineBackbone=GPT-4o-mini2025.1113.2825.61
GEPACategory=APO (GPT-4o-mini)2025.1113.2825.52
OPROCategory=APO (GPT-4o-mini)2025.1110.9426.67
GRPOBackbone=Qwen3-4B2025.1110.1623.1
BaselineBackbone=Qwen3-4B2025.116.2516.09
CoT ReasoningBackbone=Qwen3-4B2025.116.2516.25
TextGradCategory=APO (GPT-4o-mini)2025.116.2522.04
SFTBackbone=Qwen3-4B2025.115.4716.18