It seems like they all happily make up a completely incorrect answer and never simply say “I don’t know”. It seems like hallucinated answers come when there’s not a lot of information to train them on a topic. Why can’t the model recognize the low amount of training data and generate with a confidence score to determine if they’re making stuff up?
EDIT: Many people point out rightly that the LLMs themselves can’t “understand” their own response and therefore cannot determine if their answers are made up. But I guess the question includes the fact that chat services like ChatGPT already have support services like the Moderation API that evaluate the content of your query and it’s own responses for content moderation purposes, and intervene when the content violates their terms of use. So couldn’t you have another service that evaluates the LLM response for a confidence score to make this work? Perhaps I should have said “LLM chat services” instead of just LLM, but alas, I did not.
In: Technology
I suggest you read the book [On Bullshit by Harry Frankfurt](https://www.goodreads.com/book/show/385.On_Bullshit). Why? Because ChatGPT is the ultimate bullshitter, and to really understand ChatGPT, you have to understand what that means.
Bullshitters misrepresent themselves to their audience not as liars do, that is, by deliberately making false claims about what is true. In fact, bullshit need not be untrue at all. Rather, bullshitters seek to convey a certain impression of themselves without being concerned about whether anything at all is true.
ChatGPT’s training has designed it to do one thing and one thing only, produce output that the typical reader will like. Its fitness function doesn’t consider the truth of falsity of a statement. It doesn’t even know what truth or falsehood means. It boldly states things instead of saying “I don’t know” because people don’t like hearing “I don’t know” when asking a question. It expresses itself confidently with few weasel words because people don’t like to hear equivocation.
Latest Answers