Which of the following techniques uses a prompt to generate or complete subsequent prompts (streamlining the prompt development process), and to effectively guide Al model responses?

Question

Options

• AChain-of-thought prompting
• BFew-shot prompting
• CMeta prompting
• DOne-shot prompting

• A
  8% (3)
• B
  5% (2)
• C
  85% (34)
• D
  3% (1)

Explanation

Meta prompting (C) is correct because it uses a prompt to generate or refine other prompts - essentially treating the AI as a prompt engineer that produces structured, reusable prompts to guide subsequent model behavior. This self-referential technique streamlines prompt development by letting the model scaffold its own instructions.

Why the distractors are wrong:

• A (Chain-of-thought): Breaks reasoning into sequential steps to improve problem-solving, but doesn't generate new prompts.
• B (Few-shot): Provides multiple examples within a prompt to teach the model a pattern - it uses examples, not prompt generation.
• D (One-shot): Similar to few-shot but with a single example; still about demonstrating a task, not creating prompts.

Memory tip: Think of meta as "about itself" - meta prompting is a prompt about prompts, just like metadata is data about data.

No community discussion yet for this question.