today local_bar

Alisa Liu

Intern, NVIDIA

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 4

Thursday, December 4, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations

#2004 Spotlight · Brian Siyuan Zheng, Alisa Liu, Orevaoghene Ahia, Jonathan Hayase, Yejin Choi, Noah A. Smith

Language models are surprisingly robust to non-canonical tokenizations of the input, which can even lead to improved performance