Negative prompting isn’t advisable with LLMs in my experience, unless there’s support for it.
I’ve found that very intentional, descriptive and concise prompts work way better.
Used the following prompt with ChatGPT:
I’m trying to design an audio campaign for a fast broadband service. The campaign will be broadcasted to radio listeners. Can you help me come up with a script?
And got interrogatives.
Then I replied with: Let’s try to have only declarative sentences.
