Preference-Oriented Supervised Fine-Tuning: Favoring Target Model Over Aligned Large Language Models

Re-assign community