@alisynthesis @davidaugust @Robotistry funny, that’s how I formulate my own answers in conversations (what kind of answer does the user expect? 😂) But seriously, the prompt is written like a grade school math word problem. There is no grade school math test that would give your answer a correct score. And if LLMs always gave us the most pedantically accurate answer instead of the plain-reading answer they wouldn’t be useful at all.