List your company

Textual understanding on Arena-Hard

67.1Win RateUni-DPO
0.74817.97435.252.426Jun 11, 2025

Evaluation results

MethodLinks
Uni-DPOModel=Gemma-2-9B-IT2025.0667.1
SimPOModel=Gemma-2-9B-IT2025.0659.1
Uni-DPOModel=Qwen2.5-7B2025.0643.5
SFTModel=Gemma-2-9B-IT2025.0640.8
Uni-DPO v0.2Model=Llama3-8B Instruct2025.0640.6
SimPOModel=Qwen2.5-7B2025.0639.5
Uni-DPOModel=Llama3-8B Instruct2025.0637.3
SimPO v0.2Model=Llama3-8B Instruct2025.0636.5
SimPOModel=Llama3-8B Instruct2025.0633.8
DPOModel=Llama3-8B Instruct2025.0632.6
Uni-DPO v0.2Model=Llama3-8B Base2025.0630.7
SimPO v0.2Model=Llama3-8B Base2025.0629.3
SFTModel=Llama3-8B Instruct2025.0625.7
Uni-DPOModel=Llama3-8B Base2025.0623.9
SimPOModel=Llama3-8B Base2025.0623.4
DPOModel=Llama3-8B Base2025.0615.9
SFTModel=Llama3-8B Base2025.063.3