I’m sure LLMs are quickly going to learn to hallucinate (or let’s use the proper word for what they’re doing: confabulate) plausible-sounding but nonsense explanations of their thought processes at least as well as humans.