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
While an LLM can’t really think for themselves (yet), you can reduce hallucinations, if you write your prompts in a way that leaves “not knowing” a correct answer.
Example: “Give me the name of the 34th US president.” – it’s a bad prompt, because you are ordering him to spit a name and it’s likely he’ll hallucinate one if he wasn’t trained on that.
A better prompt would be: “Given your historical knowledge of US presidents, do you know the name of the 34th US president?” – it’s a good prompt, because now the LLM has room to say it doesn’t know, should that be the case.
Latest Answers