Skip to content

Interview question · LLM Application Security

What is retrieval or vector store poisoning, and how do you control it?

Poisoning a knowledge base is cheaper than poisoning a model. The answer that scores starts from who can write to the corpus and limits what a poisoned chunk can achieve.

mediumEssentialNewPreparation tags: Microsoft, Google, AmazonUpdated Nov 2026

Poisoning a knowledge base is cheaper than poisoning a model. The answer that scores starts from who can write to the corpus and limits what a poisoned chunk can achieve.

Make your first attempt

What would you say before reading the answer?

  1. 01 State the assumptions your answer depends on.
  2. 02 Explain the mechanism and choose a control.
  3. 03 Name a condition that would change your decision.

The worked answer includes an approach, scoring criteria and follow-up questions. Compare the reasoning with your own.

signing in unlocks more free answers per topic

What if · the follow-ups in this answer

The answer works through 6 follow-ups like an interviewer would

Customer support tickets are indexed so the assistant can learn from past answers and suggest fixes to agents.

  1. How can an attacker poison the assistant through tickets?
  2. What if only resolved tickets with agent replies are indexed?
  3. What if poisoned content was already indexed?
  4. What if an attacker crafts text to rank highly for common questions?
  5. What if you want early warning?
  6. What controls would you add to the ingestion pipeline?

Each rung has a nudge, a model answer, the principle that carries over and the weak answer to avoid. Sign in free to open it.

UP NEXT ON YOUR JOURNEY
DISCUSSION · 0

No comments yet. Share a correction or a different way to answer it.