promptdojo_

Tokenizers and context budget — step 3 of 7

Greedy longest-match tokenization against a 5-piece vocabulary. What does this print?