arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2604.07006cs.CL
收藏

标量多样性中的连续解释性引导

Continuous Interpretive Steering for Scalar Diversity

Sungkyunkwan University(成均馆大学)

Ye-eun Cho

AI总结:

本文提出连续解释性引导(CIS)方法,通过激活层面的连续变量探究分级语用解释,结合新数据集GraSD验证了LLM中分级语用敏感性的存在。

摘要:

语用推理本质上是分级的。不同的词汇项在语用丰富度上存在差异。标量推论通过标量多样性体现了这一特性,其中推论强度在标量项之间变化。然而,大型语言模型(LLM)中语用推理的评估通常依赖于提示式操作。除了提示层面的影响外,本研究引入了连续解释性引导(CIS),一种通过将激活层面的引导强度视为连续实验变量来探测分级语用解释的方法。为了支持这一分析,本研究引入了一个新的数据集GraSD,该数据集编码了分级标量多样性。对四个LLM的实验表明,统一的激活引导在全局上增加语用解释,但会消除项级变化;而分级激活引导则产生与标量多样性等级一致的解释变化。这表明分级敏感性编码在表示空间中,并可通过受控干预系统地恢复。共同,CIS和GraSD为评估LLM中的分级语用敏感性提供了系统性的框架。

英文摘要:

Pragmatic inference is inherently graded. Different lexical items give rise to pragmatic enrichment to different degrees. Scalar implicature exemplifies this property through scalar diversity, where implicature strength varies across scalar items. However, evaluations of pragmatic inference in large language models (LLMs) often rely on prompt-based manipulations. Beyond prompt-level effects, this study introduces Continuous Interpretive Steering (CIS), a method that probes graded pragmatic interpretation by treating activation-level steering strength as a continuous experimental variable. To support this analysis, this study introduces a new dataset, GraSD, which encodes graded scalar diversity. Experiments on four LLMs show that uniform activation steering increases pragmatic interpretations globally but collapses item-level variation, whereas graded activation steering yields differentiated interpretive shifts aligned with scalar diversity grades. It indicates that graded sensitivity is encoded in the representation space and can be systematically recovered through controlled intervention. Together, CIS and GraSD provide a principled framework for evaluating graded pragmatic sensitivity in LLMs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑