Caesar AI Atlas

Reinforcement Learning From Human Feedback (RLHF)

Also known as: Reinforcement Learning from Human Feedback (RLHF) Β· Reinforcement Learning From Human Feedback (RLHF) Β· RLHF

Caesar AI Atlas Definition

Reinforcement Learning From Human Feedback (RLHF) is a technique for aligning model behavior using human preferences or ratings. Typically, human judgments are used to train a reward model, and reinforcement learning then optimizes the generative model to produce outputs that better match those preferences.

Other Definitions

Reinforcement Learning From Human Feedback Source

A technique that involve training a "reward model" to predict how humans rate the quality of generated content, and then training a generative generative AI model to satisfy this reward model via reinforcement learning. It can be used for example to make the generative AI model more truthful or less harmful.

Reinforcement Learning From Human Feedback Source

Using feedback from human raters to improve the quality of a model's responses. For example, an RLHF mechanism can ask users to rate the quality of a model's response with a or emoji. The system can then adjust its future responses based on that feedback.

Reinforcement Learning From Human Feedback Source

A machine learning technique that trains AI models to align their responses with human preferences by using evaluative feedback.

Also Referenced In

Related Terms