You are signed in. Close

You are signed out. Close

Thank you. Your subscription has started. Close

Thank you. You will get an email when a new comic comes out. Close

We could not confirm your payment. If you paid, use the link "Sign in" at the bottom of the page. Close

Something went wrong on our side. Please try again in a moment. Close

Previous episode Attention: how an AI model works out what "it" means

How ChatGPT and other AI chatbots answer your question

Transformer: the "T" of ChatGPT, layer after layer

The "T" in ChatGPT means transformer. A transformer runs the same steps on your words again and again, in dozens of layers. Why is one pass not enough?

  1. 1 Yet the answer you get back is about watering your plant, the plant that "it" stands for.youShould I waterit more?Check thesoil first.example answer

    Yet the answer you get back is about watering your plant, the plant that "it" stands for.

  2. 2 In attention, though, "more" got only a bare "it": every token took its one look at the same moment. So how does "more" ever find out that you mean your plant?ShouldIwateritmore?I only got a bare"it". Water what?

    In attention, though, "more" got only a bare "it": every token took its one look at the same moment. So how does "more" ever find out that you mean your plant?

The rest of this episode is for subscribers

Subscribe to read every premium episode. You can cancel when you want.

Paid yearly: $99 instead of $119.88

Next episode Parameters: where an AI model keeps what it knows

Share this comic

More comics