Direct Preference Optimization
Concept
An RL-free approach that directly trains models to prefer human outputs without requiring a separate reward model.
Mentioned in 2 videos
Save the 2 videos on Direct Preference Optimization to your own pod.
Sign up free to keep building your knowledge base on Direct Preference Optimization as more episodes are added.
Videos Mentioning Direct Preference Optimization
Stanford CME295 Transformers & LLMs | Autumn 2026 | Lecture 4 - Reinforcement Learning with LLMs
Stanford Online
An approach mentioned in the context of supervised methods for RL objectives, specifically linking it to the beta coefficient in KL divergence penalties for staying close to a reference policy.
Stanford CS25: Transformers United V6 I Overview of Transformers
Stanford Online
An RL-free approach that directly trains models to prefer human outputs without requiring a separate reward model.