← Back to the AI Glossary 📖 Training & Fine-Tuning

Constitutional AI

A training technique, developed by Anthropic, where a model is guided by a written set of principles and trained to critique and revise its own outputs against them.

It's part of how Claude is trained to be helpful, honest, and harmless without needing a human to manually review every single training example.
One of 60 free AI glossary terms
Plain-language definitions for the AI jargon you'll actually run into — no email needed, ever, for this section.
Browse the Full Glossary →
More in Training & Fine-Tuning
Distillation
A technique for training a smaller, faster "student" model to mimic the behaviour of a lar...
Reward Model
A separate model trained to score how good or bad a response is, used to guide another mod...
Training Data
The text, images, or other content a model learns from during training — its entire unders...
Reinforcement Learning from Human Feedback (RLHF)
A training technique where human reviewers rate a model's responses, and the model is adju...