My understanding is that at least one form of training in the RLHF involves supplying antecedent and consequent training pairs for entailment queries. The LLM

My understanding is that at least one form of training in the RLHF involves supplying antecedent and consequent training pairs for entailment queries. The LLM seems to be only one of the many building blocks and is used to supply priors / transition probabilities that are used elsewhere in downstream part of the model.