← All methodologies

Direct Preference Optimization

DPO
Learn directly from “this answer is better.”

Train on pairs of preferred and rejected answers, directly. Standard DPO skips a separately trained reward model and the online RL practice loop.

The idea

A writing tutor gives you pairs of paragraphs and marks the stronger one. You learn from those comparisons directly, instead of first training another tutor to assign scores to new paragraphs.

Preference pair → direct update.

The task

Explain a difficult word to a beginner.

Preferred answer

Use a familiar example and one short definition.

Rejected answer

Give a circular definition with more unexplained jargon.

Feedback

A training dataset already says which answer is preferred.

DPO loss → update the model directly

Shift relative likelihood toward preferred answers, using a reference model.

A toy preference pair. Standard offline DPO trains on supplied pairs; it does not sample and score new answers inside each training step like an online RL loop.

How DPO works

  1. Collect preference pairs

    Each prompt has a preferred and a rejected answer.

  2. Keep a reference

    A reference model anchors the comparison.

  3. Train directly

    Use the DPO objective to change the relative likelihood of the two answers.

  4. Evaluate independently

    Check whether the trained model improved on new prompts.

Useful for

Preference tuning when a suitable comparison dataset is available.

The catch

DPO learns the preferences in its data, including their mistakes. Standard offline training does not automatically explore new answers.

A little more detail

Is DPO reinforcement learning?

DPO is derived from a reward-regularized preference-learning objective, but its standard training procedure avoids the usual RL rollout-and-reward loop. We label it a related preference method, not another reward source like RLHF or RLVR.

Read the research

These guides simplify the methods. The papers describe the full training recipes. Examples on this page are illustrative, not experimental results.