MEPX
Chapter 2 of 10All chapters
  1. 1What generative means
  2. 2Text generation
  3. 3Images and other media
  4. 4Working with these tools
  5. 5What to expect it to get wrong
  6. 6Customising a model
  7. 7Tools and agents
  8. 8Copyright and licensing
  9. 9Provenance and trust
  10. 10Using it well at work

Chapter 2 of 10

Text generation

Next-token prediction.

How it works

The model estimates which token is likely to come next, picks one, appends it, and repeats. Everything else, including apparent reasoning, emerges from that loop.

  • Tokens are word fragments; costs and limits are measured in them.
  • The context window is how much text it can consider at once.

Sampling

Temperature and related settings control how adventurously the next token is chosen. Low values give predictable output, high values give variety and more drift.