Data Science & ML

Hallucination

/həˌluːsɪˈneɪʃən/

Definition

When an LLM generates plausible-sounding but incorrect or fabricated information not grounded in its training data.

Example in context

"The model hallucinated a non-existent citation — always validate LLM outputs against authoritative sources."

Related terms

Practice this term

Master Hallucination in context by working through exercises in the Data Science & ML module. You'll see the term used in real engineering scenarios with multiple-choice, fill-in-the-blank, and matching drills.