#transformer

A practical introduction that breaks LLMs down into tokenization, embeddings, Transformer self-attention, pretraining, next-token prediction, inference, and context length.