ORPO-Distill: Mixed-Policy Preference Optimization for Cross-Architecture LLM Distillation

Open in new window