Prompt Leaking

Noun · AI & Machine Learning

Definitions

  1. The exposure of hidden system prompts, private instructions, or internal context through model output or prompt-injection attacks.

    In plain English: When an AI reveals hidden instructions or internal prompt text.

    Example: "Prompt leaking revealed internal policies that were never meant for end users."

Related Terms