> ## Documentation Index
> Fetch the complete documentation index at: https://docs.helicone.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Use Meta-Prompting

> Use large language models (LLMs) to create and refine prompts dynamically.

## What is Meta-Prompting

Meta-Prompting is an advanced prompt engineering method that uses large language models (LLMs) to create and refine prompts dynamically. Unlike traditional prompt engineering, Meta-Prompting guides the LLM to adapt and adjust prompts based on feedback, enabling it to handle more complex tasks and evolving contexts.

## How Meta-Prompting Works

1. Create task-specific prompts using AI
2. Guide the LLM to understand prompt structure and underlying task requirements
3. Modify prompting strategies based on context and real-time feedback
4. Work with high-level prompt design concepts
5. Instruct the LLM to evaluate and improve its prompting methods

For the official implementation, check out the paper [Meta Prompting for AI Systems](https://arxiv.org/abs/2311.11482).

## Example

**Meta-Prompt:**

> Create a prompt that will guide the LLM to analyze \[TOPIC]. This prompt should include instructions for:
>
> * Generating a clear, 3-paragraph summary
> * Identifying top 3 key arguments
> * Evaluating evidence sources
> * Suggesting 2 novel research directions. Make sure the prompt is clear and concise.

This example shows how meta-prompting can be used to create a flexible, structured approach to generating clearer prompts that can be applied across domains.

## Why use Meta-Prompting

* Versatile for a wide range of tasks
* The AI system has more autonomy in how to tackle new challenges
* More efficient resource usage and prompt optimization
* Scalability across different problem domains
* Supports AI's ongoing learning and improvement capabilities

## Tips for effective Meta-Prompting

* Define clear hierarchies and abstraction levels in your meta-prompts
* Build modular, reusable prompt components
* Test meta-prompts thoroughly across different use cases
* Follow ethical guidelines when designing prompts
* Allow for human oversight and intervention
* Regularly evaluate and update your prompting strategies

***

<Accordion title="Need more help?">
  Additional questions or feedback? Reach out to
  [help@helicone.ai](mailto:help@helicone.ai) or [schedule a
  call](https://cal.com/team/helicone/helicone-discovery) with us.
</Accordion>