Distinct social-linguistic processing between humans and large audio-language models: Evidence from model-brain alignment

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wu, Hanlin, Duan, Xufeng, Cai, Zhenguang
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911232616824832
author Wu, Hanlin
Duan, Xufeng
Cai, Zhenguang
author_facet Wu, Hanlin
Duan, Xufeng
Cai, Zhenguang
contents Voice-based AI development faces unique challenges in processing both linguistic and paralinguistic information. This study compares how large audio-language models (LALMs) and humans integrate speaker characteristics during speech comprehension, asking whether LALMs process speaker-contextualized language in ways that parallel human cognitive mechanisms. We compared two LALMs' (Qwen2-Audio and Ultravox 0.5) processing patterns with human EEG responses. Using surprisal and entropy metrics from the models, we analyzed their sensitivity to speaker-content incongruency across social stereotype violations (e.g., a man claiming to regularly get manicures) and biological knowledge violations (e.g., a man claiming to be pregnant). Results revealed that Qwen2-Audio exhibited increased surprisal for speaker-incongruent content and its surprisal values significantly predicted human N400 responses, while Ultravox 0.5 showed limited sensitivity to speaker characteristics. Importantly, neither model replicated the human-like processing distinction between social violations (eliciting N400 effects) and biological violations (eliciting P600 effects). These findings reveal both the potential and limitations of current LALMs in processing speaker-contextualized language, and suggest differences in social-linguistic processing mechanisms between humans and LALMs.
format Preprint
id arxiv_https___arxiv_org_abs_2503_19586
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Distinct social-linguistic processing between humans and large audio-language models: Evidence from model-brain alignment
Wu, Hanlin
Duan, Xufeng
Cai, Zhenguang
Computation and Language
Neurons and Cognition
Voice-based AI development faces unique challenges in processing both linguistic and paralinguistic information. This study compares how large audio-language models (LALMs) and humans integrate speaker characteristics during speech comprehension, asking whether LALMs process speaker-contextualized language in ways that parallel human cognitive mechanisms. We compared two LALMs' (Qwen2-Audio and Ultravox 0.5) processing patterns with human EEG responses. Using surprisal and entropy metrics from the models, we analyzed their sensitivity to speaker-content incongruency across social stereotype violations (e.g., a man claiming to regularly get manicures) and biological knowledge violations (e.g., a man claiming to be pregnant). Results revealed that Qwen2-Audio exhibited increased surprisal for speaker-incongruent content and its surprisal values significantly predicted human N400 responses, while Ultravox 0.5 showed limited sensitivity to speaker characteristics. Importantly, neither model replicated the human-like processing distinction between social violations (eliciting N400 effects) and biological violations (eliciting P600 effects). These findings reveal both the potential and limitations of current LALMs in processing speaker-contextualized language, and suggest differences in social-linguistic processing mechanisms between humans and LALMs.
title Distinct social-linguistic processing between humans and large audio-language models: Evidence from model-brain alignment
topic Computation and Language
Neurons and Cognition
url https://arxiv.org/abs/2503.19586