Preprint
Aug 2026
When Vocal Tone and Literal Meaning Diverge: An Acoustic-Semantic Incongruity Study for Large Audio-Language Models
CREMA-ASIS, a dataset specifically created to investigate incongruence between acoustic emotion and semantic sentiment cues, is introduced and reveals that LALMs struggle with semantic-acoustic incongruent cases, rarely predicting incongruity, and that LALMs are predominantly influenced by semantic information.
Yu-Wen Chen, William Ho, Maxim Topaz et al.
· 0 citations