Watch: How Text Becomes Numbers

Before an LLM can do anything with your prompt, it has to turn your text into numbers, because a neural network only does math on numbers. The step that does this is tokenization.

Press play. Watch a single word get split into tokens (frequent chunks of text, often subwords, not whole words), and watch each token become an integer id from the model's vocabulary. After this, the model never sees letters again, only a list of ids.