>LLMs have no concept of truth or correctness really, they are just probabilistic appending to output to match learned ideal patterns. I think they do have tho

>LLMs have no concept of truth or correctness really, they are just probabilistic appending to output to match learned ideal patterns. I think they do have those concepts, like they have many others but they are not bound to them in any way.