Understand the BasicsHow language models learn

Training data, tokens, patterns and prediction — and why learning is not the same as human understanding.

Learning patterns from text

Describe training data and repeated prediction practice at a high level.

Breaking language into tokens

Show how words and punctuation are represented in smaller pieces.

Predicting what comes next

Use a short sentence example to explain probabilities without requiring maths.

Prediction is not human understanding

Connect the mechanism to fluent answers, mistakes and hallucinations.