首页 > AI前沿 > Polarity Composition Hypothesis: Causal Evidence from Negation in NLI

Polarity Composition Hypothesis: Causal Evidence from Negation in NLI

arXiv自然语言 2026-11-18 02:30 20 阅读 查看原文

We investigate how language models process negation in natural language inference (NLI). We propose the Polarity Composition Hypothesis, a high-level causal model of this computation. Behavioral evaluations show that language-model predictions largely conform to the proposed hypothesis. Using Distributed Alignment Search, we identify causally effective subspaces aligned with the variables in the polarity-composition hypothesis. These representations transfer across negation forms and partially to a subset of MNLI. We release the dataset and code.