this is copium, the author doesn't have a good grasp on LLMs. you can't simply "ask" a language model to see if they know they're bad at math and then conclude

this is copium, the author doesn't have a good grasp on LLMs. you can't simply "ask" a language model to see if they know they're bad at math and then conclude that the response actually reflects the knowledge encapsulated in the model... sigh...