Constitutional Ai

+Constitutional AI is an approach to align [Large Language Model](/wiki/large_language_model) behavior with a set of desired principles. It leverages AI feedback, guided by a 'constitution' of rules, to refine responses and promote [AI Safety](/wiki/ai_safety) without extensive human labeling.

+## See also

+- [AI Alignment](/wiki/ai_alignment)

+- [Machine Learning](/wiki/machine_learning)

+- [Reinforcement Learning](/wiki/reinforcement_learning)