Rethinking Tokenization for Rich Morphology: The Dominance of Unigram over BPE and Morphological Alignment

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Vemula, Saketh Reddy, Dandapat, Sandipan, Sharma, Dipti Misra, Krishnamurthy, Parameswari
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2508.08424