Tag
This paper introduces TPA (Think, Plan, Ask), a proactive multi-agent dialogue framework using LLMs to systematically surface latent social language disorder traits in autism by selecting clinically grounded questioning strategies. It achieves 82.1% trait coverage, outperforming real clinical dialogues by clinicians.
This paper investigates central tendency bias in multimodal LLMs used for clinical ordinal scoring of the Clock Drawing Test, finding that LLMs compress predictions toward the middle of the scale, disproportionately affecting critical extremes. The study extends the LLM-as-judge bias literature to clinical assessment, highlighting the need for calibration-aware evaluation before deployment.