LMs trained on corpora that never show multi-modifier noun phrases still prefer English's scope-homomorphic modifier order, and noun-modifier association strength does not explain this preference.
Language Models Use Monotonicity to Assess NPI Licensing
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We investigate the semantic knowledge of language models (LMs), focusing on (1) whether these LMs create categories of linguistic environments based on their semantic monotonicity properties, and (2) whether these categories play a similar role in LMs as in human language understanding, using negative polarity item licensing as a case study. We introduce a series of experiments consisting of probing with diagnostic classifiers (DCs), linguistic acceptability tasks, as well as a novel DC ranking method that tightly connects the probing results to the inner workings of the LM. By applying our experimental pipeline to LMs trained on various filtered corpora, we are able to gain stronger insights into the semantic generalizations that are acquired by these models.
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Language Models Generalize to Human-like Word Order Preferences
LMs trained on corpora that never show multi-modifier noun phrases still prefer English's scope-homomorphic modifier order, and noun-modifier association strength does not explain this preference.