>>105959462
i still dont understand what you're trying to do
you want to feed a prompt to an llm and then use the prompt to generate, then use that same prompt you just generated with and feed it back to the llm?