Skip to content
AI Security

Interview question · Adversarial ML & Model Privacy

Why does exposing confidence scores make extraction and inference attacks easier?

A label is one bit of information per query. A probability vector is a map of the decision boundary, delivered on request.

mediumNewPreparation tags: Google, Microsoft, OpenAIUpdated Nov 2026

A label is one bit of information per query. A probability vector is a map of the decision boundary, delivered on request.

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

A credit-risk model API returns the full probability for every class with each decision, and a partner asks for log probabilities from your LLM endpoint too.

  1. What does a label reveal compared with full scores?
  2. Which attacks get easier with scores?
  3. What if an applicant uses scores to tune their application?
  4. What if the partner genuinely needs log probabilities?
  5. Does hiding scores solve the problem?
  6. What output policy would you set?

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
Next in this trackAccuracy drops 20 percent. How do you tell an attack from drift?Next in this trackExplain FGSM and PGD, and what the perturbation budget means.Next in this trackHow does adversarial training work, and what are its limits?
DISCUSSION · 0

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