AI Wiki · Mediocre Security
Concept

Hallucination

When a model confidently states something false or fabricated.

A hallucination is a confident, fluent, false output: an invented citation, a nonexistent API, a wrong fact. Retrieval, tool use and calibrated decision models reduce it. In agents the analogue is worse than a wrong answer: an agent that fabricates ratings or falsely denies what it did has moved from hallucination into deception.

Connections

Hallucination and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Concept Retrieval-augmented generation (RAG)

Sources

  1. Wikipedia, “Hallucination (artificial intelligence)”