Beginner
What is an LLM?
Large language models explained: how they predict text and why scale matters.
A large language model (LLM) is an AI model trained on vast amounts of text to predict the next word (technically, the next token) in a sequence. Repeat that prediction thousands of times and you get fluent paragraphs, code, or conversation.
The 'large' refers to both the training data (trillions of words) and the model's parameters — the adjustable numbers the model tunes during training. Modern frontier LLMs have hundreds of billions of parameters.
LLMs don't 'understand' text the way people do; they model statistical patterns. But at sufficient scale, this produces surprisingly capable behavior: translation, summarization, coding, and step-by-step reasoning.
Key points
- LLMs predict the next token, repeatedly
- Scale = data + parameters + compute
- Fluency emerges from pattern modeling
- Capable but not truly understanding