Free course · 9 lessons · 35–45 min
How LLMs actually work
What happens between pressing send and reading the reply.
What you’ll learn
Explain in plain words how a language model turns your message into a reply, how it was trained, and why it is sometimes confidently wrong.
Before you start
You've used a chatbot like ChatGPT or Claude. No math or programming is needed.
Course outline
-
Chapter 1
Words in, words out
What a language model does when it answers you, and what it actually sees when you type.
- One word at a timeA reply is built by predicting the next piece of text, over and over, with a roll of weighted dice.3 min
- Tokens, not wordsText is chopped into chunks and swapped for numbers before the model ever sees it.3 min
- Meaning as a directionEach token becomes a point in a huge space, where nearby points and shared directions carry meaning.3 min
-
Chapter 2
How it's built
The design inside the model, the training run that fills it with knowledge, and the tuning that turns it into an assistant.
- Attention: words looking at wordsHow each token pulls in information from the words before it, layer after layer.3 min
- Learning from a mountain of textBillions of random numbers become a model by guessing the next token trillions of times.3 min
- From text predictor to assistantExample conversations and human preferences teach a base model to answer instead of ramble.3 min
-
Chapter 3
Limits and open questions
Why models make things up, how tools and search help, and what researchers still can't explain.
- Why it makes things upBlurry knowledge in the weights, exact knowledge in the context window, and what happens in between.3 min
- Tools, search, and agentsA model that can only write text can still search, calculate, and act, by asking software to do it.3 min
- What nobody knows yetThe open problems: reading a model's insides, sudden abilities, and what understanding means here.3 min