tokenization

Andrej Karpathy builds the GPT tokenizer from scratch, implementing byte pair encoding along with encode and decode. He shows how many odd LLM behaviors trace back to tokenization, and covers Unicode byte encodings along the way.

Andrej Karpathy's three-hour-plus general-audience deep dive into how ChatGPT-style LLMs are built, covering the full training stack from pretraining to post-training. It also gives mental models for their quirks, like hallucination, and how to use them well.

Rob Miles explains "glitch" tokens, odd strings in large language model vocabularies that make the model produce gibberish. He covers how they likely arise from tokenizer training data, such as Reddit usernames, that never appeared in the model's own training.

Andrej Karpathy walks through how GPT assistants are trained, from tokenization and pretraining to supervised finetuning and RLHF, and gives practical prompting strategies and a view of the tooling ecosystem.