Kamus AI
Large language model
A program trained on a very large amount of text to predict how text continues.
A large language model (LLM) learns patterns from text: which words usually follow which, in what context. When you write a prompt, the model guesses the next piece of text one at a time. The result is often useful, but the model does not store facts like a database. It produces the most likely text, which is not necessarily true.
Also called: Large language model, LLM, language model, AI model
Related terms
Appears in 12 lessons
- What has actually changed
- Where AI helps, and where it does not
- AI as a highly trained text predictor
- AI as a fast new colleague
- Checking figures and spreadsheets
- The model and the app around it
- Not legal advice: AI assists, the advocate answers
- Invented provisions: never cite without opening the text
- AI does not diagnose
- Figures the model makes up
- The limits of AI in public service: it assists, it does not decide
- An agent: model, tools and a loop