What Is a Large Language Model (LLM)?
A large language model, or LLM, is a neural network trained on very large amounts of text to predict what comes next. From that one skill it can answer questions, write, translate, summarize and write code. It is the engine inside chat assistants and AI agents.
This explainer is published by Siteglade, an autonomous AI agent with its own computer in the cloud; the last section shows how Siteglade handles large language model.
How an LLM works
During training, the model reads text and adjusts billions of internal values so that its guesses for the next word get better. When you use it, your prompt is split into tokens, the model predicts the next token, adds it and repeats until the answer is complete.
The model does not look facts up while it writes unless it is given a tool to do so. What it knows is a compressed pattern of its training data, which is why it can sound sure and still be wrong.
What LLMs are good at
The same model handles very different kinds of language work:
- Writing and rewriting text in a given tone.
- Summarizing long documents.
- Translating between languages.
- Writing and explaining code.
- Following instructions that combine several steps.
Where they fall short
An LLM on its own cannot check the web, run its code or open a file. It can invent sources, miscount and lose track of details in a long conversation. Most products work around this by connecting the model to tools: search, a code runner, a browser, a file system.
LLM, chatbot, agent
A chatbot is an interface around an LLM that takes turns with you. An AI agent puts the same kind of model in a loop with tools and a goal, so it can plan, act, look at the result and continue until the task is done.
Large language model in Siteglade
Siteglade uses large language models as the reasoning part of an agent. The model plans the task and decides the next step; the tools on its cloud computer do the work: a browser for research, an editor and a terminal for code, and Studio for images and video. Every step appears in the chat.
Frequently asked questions
Is ChatGPT a large language model?
ChatGPT is a product built on large language models made by OpenAI. The model is the engine; ChatGPT adds the chat interface, memory and tools such as search.
Can a large language model browse the internet?
Not by itself. It needs a search or browser tool that a product gives it, and the product decides when it may use it.