Glossary
What is a large language model (LLM)?
A large language model, or LLM, is the kind of AI model behind assistants such as ChatGPT, Gemini and Claude. It generates text from learned patterns and the context it is given.
What a large language model is
A large language model, or LLM, is a type of AI model that works with text. Google's machine learning course describes LLMs as models that predict a token or a sequence of tokens. A token is a word, part of a word or a single character.
Given a question, an LLM produces a reply by predicting likely text, piece by piece. Google's course describes LLMs as, in essence, autocomplete mechanisms that can predict long passages.
LLMs, assistants and web access
An LLM on its own has only what it learned in training and the text it is given. Anthropic calls that text the context window, a working memory separate from the training data.
An assistant is the product built around the model, with an app, settings and tools. Web search is one such tool. Retrieval of this kind passes documents into the context window so the answer can rest on them. Anthropic calls this retrieval augmented generation.
Why LLM answers can be wrong
Google's course says LLMs hallucinate, meaning their predictions often contain mistakes. Google's Gemini privacy hub says LLM experiences can present inaccurate information as factual.
For a business, the practical risk is a wrong fact stated with confidence. The tone of an answer is no guide to whether it is right. A wrong price, address or opening time can reach a customer who never opens your website.
How Carqo relates to large language models
Because LLM answers can contain mistakes, Carqo's accuracy checks compare a fact stated in an AI answer with your own business record. The fact compared is price.
Each check uses your record as it stood when the answer was collected, so a later price change is not reported as a mistake. Prices are read from the answer by fixed rules, not by an AI model.
In practice
- Ask about your business Put customer questions about your prices, services and opening hours to the assistants your customers use. Ask each one in the same words.
- Save the answer and its sources Keep the full text, the date, the assistant and any links the answer gave.
- Compare with your record Check each stated fact against your own current information. Note which source, if any, carries the wrong detail.
Common questions
Does an LLM know about my business?
Only what it learned in training and what it reads when it is given tools such as web search. What it learned may be out of date. Ask it directly and compare the answer with your own record. Save the answer so you can compare it with later ones.
Can an LLM read my website?
It depends on the assistant and its tools. An assistant with web search can draw on pages its systems are allowed to reach. Each company documents its own crawlers and how robots.txt applies to them. Check your own robots.txt file to see which crawlers you allow.
Why does the same question get different answers?
An LLM predicts likely text, and that prediction can vary between attempts. Anthropic says identical inputs can produce different outputs. Assistants may also search the web, which brings in different sources at different times.
Does an LLM know about my business?
Only what it learned in training and what it reads when it is given tools such as web search. What it learned may be out of date. Ask it directly and compare the answer with your own record. Save the answer so you can compare it with later ones.
Can an LLM read my website?
It depends on the assistant and its tools. An assistant with web search can draw on pages its systems are allowed to reach. Each company documents its own crawlers and how robots.txt applies to them. Check your own robots.txt file to see which crawlers you allow.
Why does the same question get different answers?
An LLM predicts likely text, and that prediction can vary between attempts. Anthropic says identical inputs can produce different outputs. Assistants may also search the web, which brings in different sources at different times.
Sources
- Google machine learning crash course, What is a large language model Link checked 29 October 2026
- Anthropic, Glossary Link checked 29 October 2026
- Google, Gemini Apps Privacy Hub Link checked 29 October 2026