Tag
This paper causally analyzes how multilingual LLMs represent negation across English, French, Vietnamese, and Chinese, showing that a shared, causally controllable negation direction—and even a single sparse autoencoder feature per model—can induce or remove negation across languages and generalize to unseen ones.
This paper investigates how LLMs process negation, finding they fail to flip answers in 37–71% of negated prompts because specialized attention heads and MLP neurons suppress the original answer rather than using it to select an alternative; a confidence-weighted training objective targeting this mechanism reduces negation failures with less capability degradation.
This paper investigates how spoken negation cues in human dialogue are reflected in multimodal nonverbal behavior, using time-series classification models to distinguish negation contexts from control contexts without lexical or acoustic input.
This paper investigates how large language models handle the combination of negation and figurative language, finding that this combination poses a particular challenge and that performance depends heavily on prompt style. The authors develop new annotations for the Fig-QA dataset and analyze embedding spaces to uncover additional linguistic factors like tense and concreteness.
MIT researchers release the first multilingual negation benchmark covering seven languages and show VLMs like CLIP struggle with non-Latin scripts, while MultiCLIP and SpaceVLM offer uneven improvements across languages.