Large Language Models 101
How ChatGPT-style models actually work — and how to make them do what you want.
beginner · 45 minutes
A practical, myth-busting introduction to how large language models (LLMs) actually work. You will learn what "predicting the next token" really means, why tokenization affects cost and spelling, how training shapes behavior, and how to control and prompt models sensibly. By the end you will understand what LLMs can and cannot do *by default* — and why they sometimes confidently make things up.
What you'll learn
- What a Language Model Is: Predicting the Next Token
- Tokenization: Why a Token Is Not a Word
- Embeddings, Transformers, and Attention
- How Models Are Trained: Pretraining to RLHF
- Context Windows, Sampling, and Hallucination
- Capabilities, Limitations, and Your First Prompts