AI and LLMs · Junior · Question 47 of 75
What is a guardrail in an LLM application?
- 1A rate limit on how many users can sign up per day
- 2A check that blocks or fixes unsafe or invalid inputs and outputs
- 3A legal disclaimer shown at the bottom of every AI chat
Show the answer
Correct answer: A check that blocks or fixes unsafe or invalid inputs and outputs
Explanation
Examples: blocking personal data in outputs, validating JSON, refusing off-topic requests, or filtering harmful content.