Comment by shlewis

This is true even with the SOTA models. Making LLMs ask questions and giving answers is always a good idea. Almost every prompt I write ends with something like this: Unless undoubtedly clear, every decision and action must come from mutual agreement.