Language Model Alignment Evaluation on Arena-Hard v0.1
35.2Win Rate (%)HyPO
Evaluation Results
| Method | Links | |
|---|---|---|
| HyPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 35.2 | |
| HyPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 33.5 | |
| TR-DPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 32.5 | |
| SimPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 32.1 | |
| DPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 31.5 | |
| RainbowPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 31.3 | |
| TR-DPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 31 | |
| KTOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 30.5 | |
| FocalPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 30.2 | |
| SimPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 30.1 | |
| CPOModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 30 | |
| RainbowPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 28.6 | |
| FocalPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 27.9 | |
| HyPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 25.2 | |
| SLiC-HFModel Family=Llama-3, Model State=Instruct, Model Parameters (B)=8B2026.02 | 25.1 | |
| DPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 23 | |
| CPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 21 | |
| SimPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 20.5 | |
| FocalPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 18.7 | |
| KTOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 18.4 | |
| TR-DPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 18.2 | |
| RainbowPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 18.2 | |
| KTOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 17.5 | |
| SLiC-HFModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 16.9 | |
| DPOModel Family=Mistral, Model State=Instruct, Model Parameters (B)=7B2026.02 | 15.4 | |
| SLiC-HFModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 14.3 | |
| HyPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 13.9 | |
| CPOModel Family=Llama-3, Model State=Base, Model Parameters (B)=8B2026.02 | 12.2 | |
| SimPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 11.2 | |
| TR-DPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 9.5 | |
| RainbowPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 9.2 | |
| FocalPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 8.2 | |
| DPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 7.9 | |
| KTOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 6.6 | |
| CPOModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 6.4 | |
| SLiC-HFModel Family=Mistral, Model State=Base, Model Parameters (B)=7B2026.02 | 5.4 |