RLAIF using LLMs

RLHF but with LLMs instead of humans, to allow intelligence in open-ended tasks

Comment