Direct Preference Optimization: Your Language Model is Secretly a Reward Model

Skip the reward model and the RL — one cross-entropy loss aligns the policy directly from preferences.

Rafailov et al. · NeurIPS 2023 · Reasoning & RL. Read the paper ↗

A free, interactive, animated visual explainer of Direct Preference Optimization: Your Language Model is Secretly a Reward Model — every exhibit computed from the real formulas, with verbatim quotes from the source.

Questions

What is Direct Preference Optimization: Your Language Model is Secretly a Reward Model?

Skip the reward model and the RL — one cross-entropy loss aligns the policy directly from preferences.

Who published Direct Preference Optimization: Your Language Model is Secretly a Reward Model, and where?

Rafailov et al. — NeurIPS 2023 (arXiv:2305.18290).

Where can I find a visual explainer of Direct Preference Optimization: Your Language Model is Secretly a Reward Model?

Right here — a free, interactive, animated walkthrough of the whole paper, with exhibits computed from the real formulas and verbatim quotes from the source.

Related explainers

• DeepSeek-R1
• Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• Training language models to follow instructions with human feedback
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
• Constitutional AI: Harmlessness from AI Feedback
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
• Tree of Thoughts: Deliberate Problem Solving with Large Language Models