Concepts
Constitutional AI
Constitutional AI is the training approach Anthropic uses, where the model is given a written set of principles and taught to critique and revise its own answers against them. Rather than relying only on people rating every response one by one, the model checks itself against stated rules. Those principles are published rather than kept secret.
This entry has not yet received a source-reviewed lesson update.
Member lesson
The definition and sources are public. The complete practical lesson is for members.