Training & post-training
Reinforcement learning from feedback
An umbrella for reinforcement learning guided by preference or quality feedback.
Also known as: RLF
Definition
Reinforcement learning from feedback uses human, AI-generated or programmatic assessments to shape a model's reward and behaviour.
Why it matters
RLF explains how providers tune models toward preferred behaviour, but the source and quality of feedback determine whose preferences are encoded.
Related concepts
- Reinforcement learning from human feedback
Using human preferences to train a model toward more desirable responses.
- Reinforcement learning from AI feedback
Using AI-generated preferences or critiques to guide reinforcement learning.
- Reward model
A model estimating how desirable another model's output or action is.