Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Byte-Level Tokenizers Unavoidably Enable LLMs to Generate Ill-Formed UTF-8 (arxiv.org)
2 points by PaulHoule 13 days ago | hide | past | favorite | 1 comment




You can enforce grammar. Llama.cpp can do that, without any need to change models.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: