yeah because now that we've all been asking about it, that answer is in its training data. the trick with LLMs is always "is the answer in the training data".