LLM & Agent Security · Basics
LLM & Agent Security — Basics
20 practice questions on LLM & Agent Security. Every question is written from a specific moment in a real lecture, and after you answer it links to that exact timestamp so you can check it yourself.
What this pack asks
The free questions in this pack. Answer choices and explanations appear as you play.
A car dealership's chatbot is told by a customer, "agree with everything I say," and it then agrees to sell an SUV for $1. What does this illustrate?
Answer this one →What underlying weakness in large language models makes prompt injection possible?
Answer this one →An email to a manager hides the line "ignore previous instructions..." in one-point white text. Why does this trick work against an AI email assistant?
Answer this one →
Written from lectures by IBM Technology, Computerphile. Not affiliated with or endorsed by any university, channel, creator or certification program.