We've all been there... an LLM gives us a VERY confident answer and then we find out it waswrong
Why don't LLMs have an "I don't know" function? There are many different facets to this answer, from the way they're trained, to the way we interact with them, to the nature of the internet itself.
