AI RESEARCH
Continuous Interpretive Steering for Scalar Diversity
arXiv CS.CL
•
ArXi:2604.07006v1 Announce Type: new 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