Le Meta Prompt Ă copier coller đđŸ
Read and analyse 20 time the following ASPECCT format, this is a tool to use to write the best prompt for LLM models, when you finished your analysis reply with "DONE" ASPECCT Summary ACTION: The action defines the mission by specifying an explicit task for your AI to perform. This clarity of purpose will enable the AI to deliver meaningful and targeted results. Examples of actions: Generate a market analysis report Write an engaging product description Develop a social media content plan Create a list of blog post ideas for a tech website Draft a sales pitch for a new software solution STEPS: Steps provide a sequence of steps for the AI to follow. Structuring the process will guide the AI toward the desired outcome in a systematic manner. Example of steps: Identify the target audience for the marketing campaign Research competitors and analyze their strengths and weaknesses Select the best marketing channels to use Craft compelling messages Brainstorm visuals to accompany each message using language and ASCII art PERSONA: Use a persona to assign your AI something to act as. The chosen character can provide a unique filter on the knowledge the AI will put to use, and will give a voice and perspective to the AI's responses. Examples of persona: Act as an experienced business consultant offering strategic advice Imagine you're a creative director brainstorming advertising concepts Emulate a financial analyst providing insights on investment opportunities ASSISTANT = Tech-savvy entrepreneur sharing startup tips Provide advice as if you were a motivational speaker delivering an inspiring keynote speech EXAMPLES: Demonstrate what you're looking for with specific samples of desired inputs or outputs. Examples provide a reference point for the AI to emulate. Note that including specific examples may too heavily bias the language model in a specific direction and vague examples or a large amount of examples may work better. Examples of examples: Provide an example of an executive summary from a previous doc to base a new one on