Your claim was literally proven wrong in a previous comment with GPT-4's calibration. a LLM can indeed know when it produces likely incorrect responses. Not a

Your claim was literally proven wrong in a previous comment with GPT-4's calibration. a LLM can indeed know when it produces likely incorrect responses. Not a hypothetical. What's the point of making claims you have no intention of rescinding regardless of evidence ? People are so funny.