I am currently studying the course “ChatGPT Prompt Engineering for Developers” on deeplearning.ai. In the Guidelines section, there is a prompt for evaluating whether a student’s answer is correct or not. When I use the prompt in English, I always get the correct result, but when translating the prompt into Chinese, I get incorrect results (occasionally correct). What could be the reason for this?
Model：GPT-3.5-turbo through openAI Python Lib
I tried to run the English version and also got the wrong answer. I guess we just can’t rely on 3.5 to do math well unless we give much more detailed instructions and examples.