Glossary · LLM

What is an LLM? Explained in plain language

By Kristian Stig Henriksen · Dear Future

An LLM is a language model trained on enormous amounts of text to predict which word most likely comes next. It sounds simple, but done at sufficient scale it is what lets ChatGPT and Gemini answer questions, write text and recommend solutions in fluent language. Here is the explanation without math, and the reason it concerns your brand.

What does LLM mean?

LLM stands for large language model. “Large” refers to two things: the amount of text the model is trained on, and the number of adjustable parameters inside it, counted in billions in modern models. The architecture behind most LLMs, the transformer, was described by Google researchers in 2017 and is the foundation under practically every assistant your customers use today.

How an LLM works, briefly

During training the model reads text and practices one thing: guessing the next word. To get good at that it is forced to learn language patterns, facts, relationships and reasoning, because otherwise the next word cannot be guessed precisely. When you later ask a question, it generates its answer one word at a time from everything it has learned. That is why the answer is fluent and coherent, and why it is not a database lookup: it is a reconstruction from patterns.

That has an important consequence: a pure LLM only knows what was in its training data, and it can phrase itself just as convincingly when it is wrong. That is why the big assistants now connect the model to web search.

LLM with web search: why it can cite sources

With web search switched on, the assistant fetches current pages, reads them and writes its answer from them, with citations to the sources it used. That changes the game for brands: the answer no longer depends only on old training data, but on the pages the search finds right now. Your pages, and the sources that mention you, can move the AI answers this week, not only whenever the model is retrained.

Why LLMs matter for your brand

Your customers use LLM-driven assistants as advisors: what should I choose, what does it cost, who is best? The answer typically names a few brands, and classic search volume is falling as the assistants take over. Whether your brand is one of the names in the answers has become a measurable quantity: that measurement is what we call AI visibility.

Read on: what GEO is, the LLM SEO guide and how to measure AI visibility.

← Back to the front page