#artificial intelligence
Jun 2026
Why It Hurts: Identifying the Drivers of Negative Thoughts in Emotional Support Conversations
The AppraiSal benchmark is introduced, containing 996 emotional support conversations with human-annotated mental states, including salient cognitive appraisal dimensions, and PRISM, a multi-agent probabilistic framework grounded in Bayesian Inverse Planning, designed to improve LLMs'ability to identify context-specifi...
Hainiu Xu, ZHAOYUE SUN, Hanqi Yan et al.
· arXiv.org · 0 citations