Revisiting subword tokenization: A case study on affixal negation in large language models
This addresses tokenization issues in LLMs for linguistic tasks, but it is incremental as it focuses on a specific phenomenon.
The study measured how affixal negation affects English large language models, finding that models can generally recognize its meaning despite tokenization challenges.
In this work, we measure the impact of affixal negation on modern English large language models (LLMs). In affixal negation, the negated meaning is expressed through a negative morpheme, which is potentially challenging for LLMs as their tokenizers are often not morphologically plausible. We conduct extensive experiments using LLMs with different subword tokenization methods, which lead to several insights on the interaction between tokenization performance and negation sensitivity. Despite some interesting mismatches between tokenization accuracy and negation detection performance, we show that models can, on the whole, reliably recognize the meaning of affixal negation.