Hi, I have built a streamlit app that generates a story in the output by considering certain INPUTS from the user as given in the picture. Whenever I use the GPT-4 model, the Ai tool always generates the story, but when I use the GPT 3.5 turbo, sometimes the issues occur. When I use GPT 3.5 turbo, it sometimes gives the OUTPUT that it cannot generate the story because the reading age is not suitable (Please see attached screenshot). Now this AI tool is aimed at children, so it should always produce the story whether I use gpt-4 or gpt-3.5. GPT-4 doesn’t face this issue but GPT 3.5 does. Can someone explain to me why this happens?
GPT-3.5 and GPT-4 (and their various versions) have different capabilities and need different prompts.
A GPT-4 prompt isn’t always gonna work with GPT-3.5!
I am using the following GPT-4 and GPT 3.5 model