What Are Large Language Models (LLMs)?
Large language models power AI chatbots. Learn what tokens, context windows and training mean — and why LLMs sometimes make things up.
- Intermediate
- 11 min
- Updated
What you will learn
- Explain what an LLM is and what “large” means
- Understand tokens and context windows
- Know why LLMs hallucinate and how to reduce it
A large language model (LLM) is an AI model trained on a very large amount of text so it can understand and generate human language. AI assistants such as ChatGPT, Claude, Gemini and Copilot are built on LLMs.
Why “large”?
“Large” refers to two things: the enormous amount of text used for training, and the huge number of parameters (adjustable settings) inside the model — often billions. More scale generally means the model can capture more subtle patterns in language.
Tokens: how LLMs read
LLMs don’t read words the way we do. Text is split into small pieces called tokens — a token might be a whole word, part of a word or a punctuation mark. Usage limits and prices for AI tools are often measured in tokens.
The context window: the model’s working memory
The context window is how much text the model can consider at one time — your instructions, any documents you paste, and the conversation so far. If a conversation becomes too long, older parts may fall out of view. For long tasks, start a fresh chat and give a clear summary.
How an LLM answers
- It reads your prompt as tokens.
- It predicts the most suitable next token, then the next, and so on.
- Extra training with human feedback teaches it to be helpful, follow instructions and avoid harmful output.
Why LLMs sometimes “hallucinate”
An LLM is built to produce fluent, likely-sounding text. If it doesn’t have reliable information, it may still produce an answer that sounds right. This is called a hallucination. You can reduce it by:
- giving the model the source material and asking it to answer only from that;
- asking it to say “I don’t know” when unsure;
- asking for sources and checking them yourself;
- using tools with web search for current information.
Understanding these basics will make you a much smarter user of every AI assistant.
Practical example
Reducing mistakes: instead of asking “What is the GST rate on this product?”, paste the official notification text and ask “Based only on the text below, what rate applies? Quote the line you used.” Giving the model the source and asking it to quote reduces made-up answers.