Noticed different GPT models respond to prompt differently. Paradoxically, text-davinci-003 seems to do the best at responding to many prompts. Built a comparison tool that allows you to test how effective GPT models (text-davinci-003, gpt-3.5-turbo, and gpt-4 - if you have access to it) respond to prompts.
App: https://gpt-compare.streamlit.app
Code: GitHub - dclin/openai-model-compare: Streamlit app to test prompts on different OpenAI models