Beginner · Safety

Model Hallucinations

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. A more specific term for the phenomenon where AI models generate plausible but incorrect or fabricated information.

Technical Definition

A more specific term for the phenomenon where AI models generate plausible but incorrect or fabricated information.

How it works

Model hallucinations are a critical challenge for the reliability of generative AI, particularly LLMs. These instances occur when the model, despite being highly fluent, outputs information that is not factually grounded or exists in its training data. Addressing hallucinations is a primary focus of current AI research, employing techniques like RAG and improved model alignment.

Related Concepts

• Retrieval-Augmented Generation (RAG) — A technique that enhances LLM responses by retrieving relevant documents from an external knowledge base before generating an answer.
• Hallucination — When an AI model generates confident, plausible-sounding content that is factually incorrect or fabricated.
• Model Alignment — The process of training AI models, especially LLMs, to behave in accordance with human values, intentions, and ethical principles.