LLMs Generate Contextual Verbal-Nonverbal Mismatches in Virtual Humans
When Words and Bodies Disagree: Realizing Verbal and Nonverbal Mismatch
This study proposes a taxonomy for verbal-nonverbal mismatches and demonstrates that large language models can select appropriate mismatched behaviors like sarcasm based on social context. Human-subject evaluations confirm these cues significantly shift observers' perceptions of the virtual agent's attitude and emotional valence compared to matching baselines.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
Source：arXiv Robotics — research abstracts · arxiv.org