Databricks Certified Generative AI Engineer Associate — Question 35

A Generative Al Engineer has developed an LLM application to answer questions about internal company policies. The Generative AI Engineer must ensure that the application doesn’t hallucinate or leak confidential data.
Which approach should NOT be used to mitigate hallucination or confidential data leakage?

Answer options

• 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.

Correct answer: B

Explanation

Option B is incorrect because merely fine-tuning the model without a clear strategy does not guarantee it will avoid hallucinations or inappropriate responses. The other options provide more robust solutions: adding guardrails (A) helps filter outputs, limiting data (C) ensures users only access what they are permitted, and using a strong system prompt (D) directs the model's responses more effectively.