Showing cs.CLShow all
2 papers · 1 filter
cs.CL2023
Learn Your Tokens: Word-Pooled Tokenization for Language Modeling
Avijit Thawani, Saurabh Ghanekar, Xiaoyuan Zhu +1
Language models typically tokenize text into subwords, using a deterministic, hand-engineered heuristic of combining characters into longer surface-level strings such as 'ing' or w…
cs.CL2023★ 1 cited
Estimating Numbers without Regression
Avijit Thawani, Jay Pujara, Ashwin Kalyan
Despite recent successes in language models, their ability to represent numbers is insufficient. Humans conceptualize numbers based on their magnitudes, effectively projecting them…