EMNLP 2026 · October 28 · Budapest, Hungary

Tokenization in
Large Language Models

A tutorial and interactive tools for exploring how language models represent text.

The tutorial → Meet the presenters and read about the tutorial. Interactive playground → Explore tokenizers, compare examples, and try your own text. Visual patch renderer → Render text onto a 2D pixel strip and slice sliding-window image patches. BPE and Unigram notebook → Implement, train, and compare both algorithms in Python in your browser.