Large language model (LLM)
What is Large language model (LLM)?
A large language model (LLM) is the text brain behind ChatGPT, Gemini and Claude: it predicts the most likely next word, one word at a time.
An LLM learned from an enormous amount of text, so it writes, translates, summarises and answers fluently. On its own it does not know today's prices, your stock or a customer's order: those have to be given to it.
Large language model (LLM) in a shop
Ask an LLM "is the size M still in stock" without connecting it to your stock, and all it can do is guess.
How does Large language model (LLM) work?
- 1An LLM is trained on an enormous amount of text to predict the most likely next word.
- 2When you type a question, it reads the prompt and the conversation within its context window.
- 3It writes the answer one word at a time, each the most likely choice given what came before.
- 4For it to know your prices, stock or orders, you must give it that information, through documents (RAG) or tools.
What a shop or café can use it for
- Drafting product descriptions, posts and comment replies for a person to check and use.
- Summarising a hundred customer reviews into five main points.
- Translating product descriptions into other languages for customers abroad.
What to watch out for
- LLMs can state falsehoods confidently (hallucination): check figures, prices and policies.
- Customer data pasted into free tools may be stored or used for training: read the terms.
- An LLM's knowledge stops at a point in time; it does not know recent events unless it can look them up.
Quick questions
The LLM is the text brain. A chatbot is the program that talks to customers; modern chatbots often use an LLM inside, but some chatbots just follow a script.
They are AI assistants running on LLMs from OpenAI, Google and Anthropic, plus tools such as web search and file reading.
The large ones write Vietnamese well, but still misread abbreviations, slang and local names. A clear prompt reduces mistakes.
Keep learning beyond the glossary
Read next
- BasicsPromptA prompt is what you type to give AI a job: who it should act as, what to do and what information to use. The clearer the prompt, the closer the result.
- BasicsTokenA token is a small piece of text that AI reads and writes, such as a word, part of a word or a punctuation mark. AI prices and limits are counted in tokens.
- Data and accuracyAI hallucinationAn AI hallucination is when AI states something false in a confident voice, as if it were fact: an invented price, offer or policy.