A little-league field-lights tester is unsure how to ask for a risk-ordered condition list and first says “write me the prompt I should use.” Which prompting technique is this?
Select an answer to reveal the explanation.
Short Explanation
Meta prompting is asking the model to craft or improve the prompt — tester–AI pairing on the ask itself. Asking the model to write the prompt to use is not adding examples; it is prompting about prompting.
Full Explanation
Meta prompting uses the model to generate or refine the prompt itself—exactly when a little-league field-lights tester is unsure how to ask for a risk-ordered condition list and first says “write me the prompt I should use.” Civic teams can meta-prompt to draft a better ask, then human-review and execute that prompt against the lights story. Meta is about improving the prompt artefact, not yet treating the final condition list as the only goal of the first message.
Few-shot is multiple examples of desired output; requesting a prompt is not multiple exemplars of risk-ordered conditions. Zero-shot is not automatic for any request that mentions prompts. One-shot is not triggered merely by the phrase “write me the prompt,” which is not a worked input→output example of the condition list.
Exam caveat: meta prompting = model helps write or refine the prompt. Separate it from shot-count techniques and from chaining of work products.
Operational check: ask the model for a draft prompt aimed at risk-ordered field-lights conditions; edit for Role, Context, Instruction, Constraints, and Output format; execute the revised prompt; store both the meta step and the final prompt in the little-league municipal library.