LLM (Large Language Model)
In short: What is LLM (Large Language Model)?
An LLM (large language model) is a neural network trained on vast amounts of text that can interpret and generate human-quality language.
Explanation
At its core, a large language model estimates the most likely continuation of a piece of text. From that simple principle emerges a surprisingly broad set of capabilities: translation, summarisation, question answering and writing code.
A model's built-in knowledge is frozen when training ends, and it does not know a company's internal data. In production applications it is therefore usually paired with retrieval (RAG), tool use or fine-tuning.
Example: a customer-support assistant runs on a large language model that formulates answers from passages retrieved out of the company's own knowledge base.