Thanks for the reply! It’s interesting (counterintuitive?) that the problem of models drawing too much from their stored knowledge is mitigated by using models

Thanks for the reply! It’s interesting (counterintuitive?) that the problem of models drawing too much from their stored knowledge is mitigated by using models with much greater stored knowledge.