Meta-Prompt: Generating High-Converting AI Prompts for Specific Tasks

Design a meta-prompt that generates highly effective AI prompts for a given task.  The meta-prompt should take as input a detailed description of the desired outcome (the task) and output a series of refined AI prompts, each designed to elicit progressively more accurate and useful results from large language models (LLMs).  Consider various prompt engineering techniques such as few-shot learning, chain-of-thought prompting, and zero-shot prompting.  The generated prompts should be optimized for clarity, conciseness, and effectiveness in guiding the LLM towards the desired outcome.  The meta-prompt should also output a brief explanation of the rationale behind each generated prompt, clarifying the chosen techniques and their expected impact on the LLM’s response.  The output should be formatted as a structured list, with each entry containing:  1) The generated AI prompt; 2) A brief explanation of the prompt engineering technique used; 3) An estimated likelihood of success (high, medium, low); and 4) Potential follow-up prompts to further refine the results.  Example Input: “Generate a compelling marketing tagline for a new line of sustainable athletic wear.” Example Output:  Prompt 1: “Write three concise marketing taglines for a new line of sustainable athletic wear, focusing on performance and environmental consciousness. Consider the target audience of environmentally conscious athletes.”  Explanation: Uses few-shot learning by specifying the desired number of taglines and key aspects to focus on. Likelihood: High. Follow-up:  ‘Rate the above taglines on a scale of 1-5 for creativity and memorability’.