Which statement correctly describes how to mitigate prompt gaps and leakage?

• A. Rely on verbose prompts and ignore safety checks.
• B. Increase model size to compensate the lack of context.
• C. Provide explicit context, use checklists, constraint handoffs, and content filtering.
• D. Increase the temperature to reduce risk.

Answer: (C)

Mitigating prompt gaps and leakage works best when you provide explicit context, use checklists, perform constraint handoffs, and apply content filtering.

Explicit context sets clear boundaries for what the model should consider and what it should avoid. By outlining allowed topics, sensitive constraints, and the desired level of detail, you reduce the chances the model fills gaps with unintended or unsafe information. Checklists give you a repeatable way to verify that safety steps are followed before the model proceeds, catching potential issues in a structured way rather than relying on memory or chance. Constraint handoffs make sure any downstream processes or systems receive the same safety requirements, so a permissive prompt doesn’t slip through the cracks when parts of the pipeline interact. Content filtering actively screens outputs against policy and risk signals, providing an additional gate to prevent leakage before anything is shown to users.

Why the other approaches aren’t as effective: simply making prompts longer or more verbose without enforcing safety checks doesn’t guarantee that the model won’t reveal or generate inappropriate content. Increasing model size doesn’t automatically fill in missing safety context or enforce constraints; it can even magnify hidden biases or gaps. Raising the temperature makes outputs more random and less predictable, which can increase the risk of unsafe or leaking content rather than reducing it.

So, combining explicit context, structured verification, clear constraint transfers, and robust filtering gives you concrete controls to minimize prompt gaps and leakage.