Loading slide

What a Trained Model Actually Is

  1. 01What a Trained Model Actually Is
  2. 02Memory card
  3. 03A model is a file
  4. 04Two files
  5. 05How the two fit together
  6. 06A Piano
  7. 07The valuable file
  8. 08You connect to it
  9. 09Where does training data come from?
  10. 10Cleaning the data
  11. 11The network can't read
  12. 12Guess the next word
  13. 13Right or wrong
  14. 14One nudge, a billion times
  15. 15Take a breath
  16. 16Where your memories live
  17. 17A base model is not an assistant
  18. 18Why it's called pre-training
  19. 19The weights are the model
  20. 20The scale of it
  21. 21Why so many?
  22. 22What a weight actually is
  23. 23Everywhere and nowhere
  24. 24Using the model: inference
  25. 25The whole arrangement
  26. 26Reinforce your understanding
  27. 27Question: What is a model, really?
  28. 28Question: Learning from raw text
  29. 29Question: The model doesn't know what it doesn't know
  30. 30Quiz: answer
  31. 31The machine learned something real.
  32. 32Want to go deeper?
21 / 31
BackNext

Why so many?

Think of painting a face.

With a handful of brushstrokes, you get something roughly face-shaped. With thousands, you can capture a real expression: the exact set of a mouth, the hint of a smile that might be warm or might be mocking.

Parameters work the same way. Each one is a tiny adjustment. A few capture rough patterns. Enough of them, and the model can hold fine distinctions, the difference between "Great job" said plainly and "Great job" meant as a dig.

No single parameter holds "sarcasm," just as no single brushstroke holds "a smile." The nuance lives in how they all combine, the same spread we just saw.

More of them is not a promise of a better model. But it raises the ceiling on how much nuance one can hold.

So that is the case for billions. But what is a single one, up close?

Citations
Citations