Token visualizer

I was planning to give a Product Knowledge session on LLMs. My initial plan was to show the Tiktokenizer tool (tiktokenizer.vercel.app) while discussing tokens. Then I had an idea to build something for my "vibecoding" project, which led to the creation of this cl100k visualizer.

To make the session more interesting, I also added some tricky sentences as N-gram examples to go along with the tokenization.

LLM tokenization demo screenshot