Skip to content

Glossary: Vulnerability, Asset and Change Management

LLM (large language model)

Also written: LLM, large language model, LLM app

An AI model trained on huge amounts of text that writes text in reply to a prompt, like the one behind Kettle & Crane's customer service assistant. Treat an LLM app as software with permissions: limit its data and tools, and don't trust its output.

Where the book teaches it

Related terms

Prompt injectionExcessive agencyOWASP Top 10 for LLM Applications