A large language model is a probability distribution over the next token, fit to a corpus the size of the readable internet. Everything it knows is encoded as a tendency: which words, in which order, are likely to follow which.
In plain language
An LLM is not a search engine and not a database — it is a probability machine for language. Given the words you have written, it ranks every possible next word and picks among the likely ones. Because it has read so much, the choices it makes usually look fluent and often look correct. They are not always correct, which is the part beginners most often miss.
An everyday picture
Think of Large Language Model less like a thinking person and more like someone who has read an enormous amount and now finishes other people's sentences for a living. They have absorbed the shape of the work; they have not memorised any one page.
Where it shows up
Large Language Model tends to sit inside products that need to read, write, or recognise without a hard-coded rule — assistants, search, document tools, voice apps. It is rarely the only moving part, but it is often the part the user feels.
A small example
ChatGPT, Claude, and Gemini are all LLMs. When you type 'write me a polite email declining a meeting,' the model predicts, word by word, the most likely polite-email continuation.
Common misunderstanding
One line to take with you
Large Language Model is statistics worn well. Useful for patterns; double-check it for facts.
Frequently asked
Because it is optimising for plausible-sounding language, not truth. When the training data is thin or contradictory, the most likely sentence may not match reality.