Free Courses on How Large Language Models Work
7 free courses and videos, grouped by the level each provider gives, with the prerequisites and certificates their pages state.
What is a large language model?
A large language model (LLM) is the kind of AI behind ChatGPT, Claude and Gemini: a program that writes by predicting the next small piece of text, trained on a very large amount of writing. Underneath it is a neural network built on a design called the transformer, which uses a mechanism called attention to work out which earlier words matter.
The resources below explain that from the ground up. The visual videos need no coding; the deeper courses build the pieces in Python.
Who these courses are for
Anyone curious how chatbots work. The video explainers need no coding; Hugging Face's course and Karpathy's lectures expect Python.
Beginner courses (5)
But What Is a Neural Network?
3Blue1Brown
Chapter 1 of 3Blue1Brown's deep learning series: what a neural network is, explained visually.
Transformers, the Tech Behind LLMs
3Blue1Brown
Chapter 5 of the deep learning series: how a transformer turns text into a prediction of what comes next.
Attention in Transformers, Step by Step
3Blue1Brown
Chapter 6 of the deep learning series: the attention mechanism inside a transformer, worked through visually.
Intro to Large Language Models
Andrej Karpathy
A one-hour general-audience talk on what large language models are and how they are built.
Deep Dive into LLMs like ChatGPT
Andrej Karpathy
A long-form walk through how chatbots such as ChatGPT are built, from pre-training data to the assistant.
Intermediate courses (1)
Advanced courses (1)
View and sign-up counts are the providers’ own figures, read 18 October 2026.
Frequently asked questions
Where can I learn how LLMs work for free?+
Free options include 3Blue1Brown's deep learning videos, which explain neural networks, transformers and attention visually; Andrej Karpathy's talks and lectures; and Hugging Face's LLM Course.
Do I need maths or coding to understand LLMs?+
Not for the overview. The 3Blue1Brown videos and Karpathy's one-hour talk explain the ideas without code. Hugging Face states its LLM Course requires a good knowledge of Python, and Karpathy's Zero to Hero series builds everything in code.
What is a transformer in AI?+
A transformer is the neural-network design used by almost all large language models, introduced in the 2017 paper Attention Is All You Need. Its key mechanism, attention, lets the model weigh which earlier words in the text matter for each new word.