Term
Meta-Prompting
Meta-prompting is the technique of having an LLM write or improve a prompt for a task rather than formulating it yourself — the prompt becomes the output.
Meta-Prompting — explained in more detail
Instead of crafting the perfect prompt yourself, you let the model design one: “write a system prompt that models a code reviewer for Python web applications, with clear format rules and three few-shot examples.” The LLM draws on its model knowledge about proven prompt structures — role, context, instructions, examples, output format — and produces a ready-to-use prompt. That prompt is then used (often with the same model) for the actual task. OpenAI and Anthropic now offer such generators directly inside their playgrounds as “prompt improver” features.
Example / Practical context
A typical loop: submit a weak prompt, ask the model “improve this prompt for clarity, specificity and robustness”, test the result, iterate if needed. Especially valuable for complex agent system prompts whose manual upkeep quickly becomes unwieldy. The limit: the model can only suggest structures it has seen in training — genuinely novel prompt architectures still need human invention.
Distinction from related concepts
Prompt engineering is the manual version of the same activity. Prompt templating uses fixed templates with placeholders and is more of an engineering than an AI task. Self-Refine has the model improve its own answer, not the prompt — the output is the answer, not the prompt.
Entdecke mehr
Effort level and deep thinking: two independent axes for AI tasks
Effort scales breadth, deep thinking scales depth. When each setting makes sense — with three clear examples and one rule of thumb.
GlossarChain-of-Thought
Chain-of-Thought (CoT) is a prompting technique that asks the model to spell out its reasoning in intermediate steps — boosting accuracy on multi-step tasks.
LexikonPrompt Security — Injection, Leaking, Guardrails
How prompt injection, prompt leaking, and jailbreaks work — and which defenses (guardrails, spotlighting, sanitization) actually help.