List your company

MT-Bench on UltraMedical Preference

6.9MT ScoreSSPO
5.1325.5916.056.509Oct 28, 2025

Evaluation results

MethodLinks
SSPOBackbone=Llama3 (8B), Size=10%2025.106.9
SSPOBackbone=Llama3 (8B), Size=1%2025.106.7
DPOBackbone=Llama3 (8B), Size=1%2025.106.5
SimPOBackbone=Llama3 (8B), Size=10%2025.106.5
DPOBackbone=Llama3 (8B), Size=10%2025.106.4
ORPOBackbone=Llama3 (8B), Size=1%2025.106.4
ORPOBackbone=Llama3 (8B), Size=10%2025.106.4
SimPOBackbone=Llama3 (8B), Size=1%2025.106.4
SSRMBackbone=Llama3 (8B), Size=1%2025.106.4
SSRMBackbone=Llama3 (8B), Size=10%2025.106.4
KTOBackbone=Llama3 (8B), Size=1%2025.106.4
KTOBackbone=Llama3 (8B), Size=10%2025.106.4
SPABackbone=Llama3 (8B), Size=1%2025.106.4
SPABackbone=Llama3 (8B), Size=10%2025.106.4
DPOBackbone=Mistral (7B), Size=10%2025.105.3
SimPOBackbone=Mistral (7B), Size=1%2025.105.3
SimPOBackbone=Mistral (7B), Size=10%2025.105.3
SSRMBackbone=Mistral (7B), Size=10%2025.105.3
KTOBackbone=Mistral (7B), Size=10%2025.105.3
SPABackbone=Mistral (7B), Size=10%2025.105.3
SSPOBackbone=Mistral (7B), Size=10%2025.105.3
DPOBackbone=Mistral (7B), Size=1%2025.105.2
ORPOBackbone=Mistral (7B), Size=1%2025.105.2
ORPOBackbone=Mistral (7B), Size=10%2025.105.2
SSRMBackbone=Mistral (7B), Size=1%2025.105.2
KTOBackbone=Mistral (7B), Size=1%2025.105.2
SPABackbone=Mistral (7B), Size=1%2025.105.2
SSPOBackbone=Mistral (7B), Size=1%2025.105.2