Quiz
Ops
Features
Pricing
Security
About
← Back to home
Question 1 of 15
20:00
What is a direct prompt injection attack?
A method for improving the model's factual accuracy
A network-level attack that intercepts API traffic between the client and the LLM provider
An attacker directly supplies malicious instructions in their own input to the model, attempting to override the application's system prompt or intended behavior
A technique for compressing prompts to save on token costs