AI & Machine Learning · reviewed in July 2026
Large language model (LLM)
A large language model (LLM) is a neural network trained on huge amounts of text to predict the next word (or token) in a sequence, and which as a result gains the ability to generate coherent text, answer questions, translate, and follow instructions. GPT and Llama are well-known examples.
Frequently asked questions
How does an LLM 'understand' language?
It doesn't understand the way a person does: it learns statistical patterns of which words tend to follow others, from trillions of text examples, and uses those patterns to generate plausible continuations.
What is a token?
It's the smallest unit of text an LLM processes — sometimes a whole word, sometimes part of one. The model processes and generates text token by token, not necessarily letter by letter or word by word.
What is a hallucination in an LLM?
It's when the model generates information that sounds convincing but is false or made up, because its goal is to produce plausible text, not to verify facts against a trustworthy source.