I want to prove to someone that ai won't ever say "I don't know" because it's just a bunch of numbers predicting other numbers, so i am looking for some questions which i can ask a llm for which the llms reply with an obviously incorrect answer instead of admitting that they don't know