To write good prompts, it helps to know what a language model actually does with one. At its core, an LLM is a function that reads text and predicts the next token, over and over, until it has produced a reply. It does not look anything up or follow your instructions the way a program would. It continues the text in the most fitting way, based on patterns learned from huge amounts of writing.
The lessons start with a short overview of what a language model is and how training shapes it, then a longer tour of how models are tuned into assistants, how they use tools, and where prompt injection comes from. They finish with practitioners explaining what makes a good prompt and why wording matters so much.
Watch the lessons in order, then continue on to Writing Clear Instructions.
A short animated explainer of what a language model is: a function that predicts the next word, shaped by training on huge amounts of text.
Video 8minA general-audience tour of how LLMs work, how they are trained and tuned into assistants, how they use tools, and where jailbreaks and prompt injection come from.
Video 60minAnthropic's prompt engineers discuss what makes a good prompt, how to refine one through iteration, and how prompting is changing as models improve.
Video 1h 16min