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?
14 / 31
BackNext

One nudge, a billion times

Now we act on that error. Using backpropagation, the method we covered earlier, it travels backwards through the entire network. Every weight that contributed to the wrong guess gets a small share of the blame, proportional to how much influence it had. Each one gets nudged, just slightly, in the direction that would have made "stars" a little more likely next time. The bigger the gap, the bigger the nudge.

Not dramatically. Barely perceptibly. A single example changes almost nothing.

But this doesn't happen once. The model works through the entire dataset, one piece of text after another: guess, uncover the answer, measure the gap, nudge. It is the same loop that turned the Perceptron's dials sixty years ago, just pointed at words now, and run billions of times over. And then it does the whole thing again, looping over the data several times. Each pass, the weights are a little sharper, so the same sentence gets guessed a little better than before.

Each tiny nudge accumulates. Drag the slider to watch it happen: at the start, "stars" is buried in the noise, no more likely than "banana." As the examples pile up, the weights settle, and slowly that one bar climbs above all the others. The noise collapses into a confident guess.

After enough of this, something has changed. The weights are no longer random. They contain something.

Citations(2)↓
  1. 1. nature.com
  2. 2. developers.google.com
Citations(2)↓
  1. 1. nature.com
  2. 2. developers.google.com