AI Fundamentals

RLAIF

Reinforcement Learning from AI Feedback — like RLHF but the preference labels come from another LLM instead of humans. Complements Constitutional AI; substantially cheaper than human labelling at the cost of inheriting the labeller model's biases.

Related terms

Next steps

Where this fits

RLAIF is part of the AI Fundamentals vocabulary used in the Generative AI Maturity Framework. See the full glossary for the complete set of 149 defined terms, or take the free maturity assessment to see where your organisation stands.