What is an LLM?
An LLM (large language model) is an AI model trained on huge amounts of text that learns to predict which word, or part of a word, comes next. That ability powers tools that write, summarize, translate and answer questions, such as Google's Gemini and GPT, the model family behind ChatGPT.
Updated:
How an LLM works
A language model estimates the probability of a unit of text appearing within a longer sequence. That unit is called a token, and it can be a word, part of a word or even a single character.
Google's machine learning glossary defines an LLM as a language model with a very high number of parameters. In everyday use, it means any Transformer-based language model, such as Gemini or GPT. The Transformer architecture was developed at Google and introduced in the June 2017 paper "Attention Is All You Need".
An example
When you ask ChatGPT to "write a post about hiring", the model doesn't retrieve a finished post. It generates the text token by token, based on what it learned fits next. That is why two identical requests can give two different answers.
Common confusion: LLM vs chatbot
ChatGPT and Gemini are products. The LLM is the engine inside them. The product adds a chat interface, conversation memory and sometimes web search. Without search, the model answers only from what it learned in training, so recent information may be missing.
How it relates to Velut
Velut has two separate things. For managed clients, I write by hand, in their voice. Separately, Velut AI is a product: a personal AI model that learns from your own LinkedIn writing, for people who prefer to write themselves. And when an LLM answers with search, it picks sources. That is the basis of GEO.
Frequently asked questions
What is the difference between an LLM and ChatGPT?
The LLM is the model. ChatGPT is an OpenAI product built on GPT models that adds a chat interface and other capabilities. The same goes for Google's Gemini: there is a model, and there is an app.
What is a token?
A token is the small unit of text a model works with. It can be a word, part of a word or a single character.
Does an LLM know what is true?
Not necessarily. It predicts plausible text, so it can write a sentence that sounds confident but is wrong. Always check facts against a source.