AlpacaEval 2.0 on UltraMedical Preference
17.7LCSSPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SSPOBackbone=Llama3 (8B), Size=10%2025.10 | 17.7 | 18.4 | |
| SPABackbone=Llama3 (8B), Size=10%2025.10 | 15.7 | 16.9 | |
| KTOBackbone=Llama3 (8B), Size=10%2025.10 | 14.2 | 15.2 | |
| SSRMBackbone=Mistral (7B), Size=10%2025.10 | 13.1 | 14.6 | |
| SSRMBackbone=Llama3 (8B), Size=10%2025.10 | 12.9 | 15.8 | |
| SSPOBackbone=Mistral (7B), Size=10%2025.10 | 12 | 17 | |
| SPABackbone=Mistral (7B), Size=10%2025.10 | 11.3 | 16.2 | |
| DPOBackbone=Mistral (7B), Size=10%2025.10 | 11.2 | 7.2 | |
| KTOBackbone=Mistral (7B), Size=10%2025.10 | 11.1 | 15.6 | |
| SimPOBackbone=Llama3 (8B), Size=10%2025.10 | 10.5 | 4.5 | |
| SimPOBackbone=Mistral (7B), Size=10%2025.10 | 10.2 | 7.6 | |
| DPOBackbone=Mistral (7B), Size=1%2025.10 | 8.7 | 4.8 | |
| DPOBackbone=Llama3 (8B), Size=10%2025.10 | 7.8 | 6.2 | |
| SSPOBackbone=Mistral (7B), Size=1%2025.10 | 7.6 | 13.2 | |
| ORPOBackbone=Llama3 (8B), Size=10%2025.10 | 7.5 | 5.6 | |
| SPABackbone=Mistral (7B), Size=1%2025.10 | 6.7 | 10.1 | |
| ORPOBackbone=Mistral (7B), Size=10%2025.10 | 6.5 | 5 | |
| KTOBackbone=Mistral (7B), Size=1%2025.10 | 6.4 | 9.8 | |
| SimPOBackbone=Mistral (7B), Size=1%2025.10 | 6 | 7.2 | |
| SSRMBackbone=Mistral (7B), Size=1%2025.10 | 6 | 5.9 | |
| SSPOBackbone=Llama3 (8B), Size=1%2025.10 | 5.1 | 6.7 | |
| SPABackbone=Llama3 (8B), Size=1%2025.10 | 4.3 | 5.8 | |
| SSRMBackbone=Llama3 (8B), Size=1%2025.10 | 3.9 | 5 | |
| ORPOBackbone=Mistral (7B), Size=1%2025.10 | 3.7 | 3.1 | |
| KTOBackbone=Llama3 (8B), Size=1%2025.10 | 3.7 | 5.4 | |
| SimPOBackbone=Llama3 (8B), Size=1%2025.10 | 2.8 | 3.8 | |
| DPOBackbone=Llama3 (8B), Size=1%2025.10 | 2.6 | 5.3 | |
| ORPOBackbone=Llama3 (8B), Size=1%2025.10 | 2.2 | 4.8 |