News for hackers. Stuff that matters.
Re:The Bitter Lesson is coming for Tokenization by search_facility on Wednesday June 25, 2025 @05:15PM (#44381905)
turns out - no, by intuition they should do this for sure - but no.

UPD: Found the paper: - https://huggingface.co/papers/2502.09741 - https://fouriernumber.github.io/

in paper mentioned “number” is a single sort-of “token” with numeric value, so network dealing with numbers like real numbers, separately from char representation. All the math happens directly on “number value”. In majority of current models numbers are handled like sequences of chars

There are no comments on this comment yet.