Would love to see how LLMs fare on this type of study. Paradoxically, it may be comforting to know they are at most as overconfident as humans in situations wit

Would love to see how LLMs fare on this type of study. Paradoxically, it may be comforting to know they are at most as overconfident as humans in situations with limited information. Perhaps an overconfidence scale on which we assess models as a proxy for hallucination proclivity.