61
Dominated by popular culture, brands, platform-speak, and non-words. Tokenization is, therefore, a political act: it defines what can be represented and how the world becomes computationally representable.
Welcome to 1984
This was a very interesting talk. Thanks for sharing it.
Why 51? If you assign a binary number to a split, so g|i|r|af|fe is 111010 and gira|f|fe is 000110, that would be 2^7-1. I didn't get how this is counted.
It doesn't explain tokenizers well (or at all). There are better videos on the subject.
Anyway. Suppose you wanted to spell giraffe with the English alphabet in any arbitrary, phonetic way. You could also spell, for example; "jeeruff", or "djirough". You could count how many phonetically correct ways there are to spell "giraffe".
Tokenizers break a text into sequences of characters (even individual characters), called tokens. Different tokenizers use different tokens. The one they use here has "gira" as a token, but also "g", "i", "r", and "a". So you could tokenize the same text in different ways. They have a slide where they show the possibilities.
this post was submitted on 16 May 2026
61 points (100.0% liked)
Technology
84769 readers
3261 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS