List your company

Response Generation on UltraFeedback (val)

88.1BERTScoreRPO
78.42880.93983.4585.961Jun 16, 2025

Evaluation results

MethodLinks
RPOModel=Mistral-7B2025.0688.134.80.791108——
RPOModel=Qwen2.5-7B2025.068834.60.791107——
RPOModel=LLaMA-3-8B2025.0687.934.40.790106——
RPOModel=FlanT5-Large2025.0687.131.60.98684——
RPOModel=FlanT5-XL2025.068729.70.88891——
DROModel=Mistral-7B2025.0686.932.61.1849488.986.5
DROModel=Qwen2.5-7B2025.0686.832.31.1849589.887.4
DROModel=LLaMA-3-8B2025.0686.7321.2839389.586.9
RPOModel=FlanT5-Small2025.0686.628.91.28372——
KTOModel=Mistral-7B2025.0686.431.91.2829189.887.3
KTOModel=Qwen2.5-7B2025.0686.331.51.2828990.788.2
KTOModel=LLaMA-3-8B2025.0686.231.11.3818890.287.9
SFTModel=Mistral-7B2025.0686.131.21.3818888.686.2
SFTModel=Qwen2.5-7B2025.068630.91.3808689.487.1
DROModel=FlanT5-Large2025.0685.930.31.3797185.983.8
SFTModel=LLaMA-3-8B2025.0685.830.51.4808589.186.8
DROModel=FlanT5-Small2025.0685.625.21.5766384.582.1
DROModel=FlanT5-XL2025.0685.4251.2787385.183
SFTModel=FlanT5-XL2025.0685.124.41.6746686.284.1
KTOModel=FlanT5-Small2025.068523.71.3745983.180.8
SFTModel=FlanT5-Large2025.0684.624.11.7736085.883.2
SFTModel=FlanT5-Small2025.0684.222.81.8715481.479.2
KTOModel=FlanT5-Large2025.068424.92.475658784.6
KTOModel=FlanT5-XL2025.0678.824.82.4705792.790.1