also you shouldn't have to specify it in your prompt, could confuse the llm they process tokens simultaneously unlike a human