How AI Actually Works: From Neurons to ChatGPT
Unit 3 · Lesson 2

Thumbs Up, Thumbs Down, and the Hidden Instructions

You'll be able to describe how human preference ratings and a system prompt shape a model's tone and limits, and decide which of the three stages is responsible for fixing a given misbehaviour.
Next: Why "It" Means the Suitcase: Attention Scores in Plain Words →
0:00 / 10:21
Chapter 1 of 5 — Four Boxes Between Raw and Helpful
Like it? Keep going.

A free account saves your progress in the free library. A paid plan puts a tutor right here: pause and ask about anything on screen. It adds flashcards and practice tests, and lessons built or rebuilt for the way you learn.

Sign up free