AlpacaEval 2.0 on DSP Business
17.9LCSSPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SSPOBackbone=Mistral (7B), Size=10%2025.10 | 17.9 | 8.8 | |
| SSPOBackbone=Mistral (7B), Size=1%2025.10 | 17.2 | 7.1 | |
| KTOBackbone=Mistral (7B), Size=10%2025.10 | 16.7 | 7.5 | |
| DPOBackbone=Mistral (7B), Size=10%2025.10 | 16 | 6.8 | |
| ORPOBackbone=Mistral (7B), Size=10%2025.10 | 16 | 7 | |
| SPABackbone=Mistral (7B), Size=10%2025.10 | 16 | 7.2 | |
| ORPOBackbone=Mistral (7B), Size=1%2025.10 | 15.9 | 6.5 | |
| SimPOBackbone=Mistral (7B), Size=10%2025.10 | 15.9 | 7.1 | |
| SSRMBackbone=Mistral (7B), Size=10%2025.10 | 15.9 | 6.9 | |
| KTOBackbone=Mistral (7B), Size=1%2025.10 | 15.7 | 6.8 | |
| SimPOBackbone=Mistral (7B), Size=1%2025.10 | 15.5 | 6.5 | |
| SPABackbone=Mistral (7B), Size=1%2025.10 | 15.4 | 6.5 | |
| SSRMBackbone=Mistral (7B), Size=1%2025.10 | 15.2 | 6.7 | |
| DPOBackbone=Mistral (7B), Size=1%2025.10 | 15 | 6.5 | |
| SSPOBackbone=Llama3 (8B), Size=10%2025.10 | 5.7 | 5.7 | |
| ORPOBackbone=Llama3 (8B), Size=10%2025.10 | 4.6 | 4.7 | |
| KTOBackbone=Llama3 (8B), Size=10%2025.10 | 4.4 | 4.6 | |
| SPABackbone=Llama3 (8B), Size=10%2025.10 | 4.3 | 4.1 | |
| SSRMBackbone=Llama3 (8B), Size=10%2025.10 | 3.8 | 3.5 | |
| DPOBackbone=Llama3 (8B), Size=10%2025.10 | 3.7 | 3.6 | |
| SSPOBackbone=Llama3 (8B), Size=1%2025.10 | 3.7 | 3.6 | |
| ORPOBackbone=Llama3 (8B), Size=1%2025.10 | 3.6 | 3.1 | |
| SimPOBackbone=Llama3 (8B), Size=10%2025.10 | 3.6 | 3.5 | |
| KTOBackbone=Llama3 (8B), Size=1%2025.10 | 3.2 | 3.1 | |
| SSRMBackbone=Llama3 (8B), Size=1%2025.10 | 3.1 | 2.9 | |
| SPABackbone=Llama3 (8B), Size=1%2025.10 | 3 | 2.7 | |
| DPOBackbone=Llama3 (8B), Size=1%2025.10 | 2.7 | 2.1 | |
| SimPOBackbone=Llama3 (8B), Size=1%2025.10 | 2.7 | 2.1 |