Certified Generative AI Engineer AssociateDemo

By Databricks · Question Mode

QuestionQ66

Governance

A Generative AI Engineer has developed an LLM application that answers questions about internal company policies. The Generative AI Engineer must ensure the application does not hallucinate or expose confidential data.

Which approach should NOT be used to mitigate hallucination or confidential-data leakage?

  • A Add guardrails to filter outputs from the LLM before it is shown to the user
  • B Fine-tune the model on your data, hoping it will learn what is appropriate and not
  • C Limit the data available based on the user’s access level
  • D Use a strong system prompt to ensure the model aligns with your needs.
Explanation

Fine-tuning is not a dependable security or authorization control: it can embed sensitive training data in model behavior and does not reliably prevent inappropriate disclosure. Confidential information should instead be controlled through access-aware data retrieval and output safeguards; guardrails can filter sensitive information and detect responses that are not grounded in source material.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!