New benchmark and PRISM framework improve LLMs' ability to identify emotional appraisal dimensions

Researchers have introduced the AppraiSal benchmark, a dataset of 996 emotional support conversations annotated with human mental states and salient cognitive appraisal dimensions. This benchmark aims to help Large Language Models (LLMs) better infer context-specific appraisal dimensions, which are crucial for modifying cognitive appraisals in emotional support tasks. To address this, a new framework called PRISM, based on Bayesian Inverse Planning, has been developed, demonstrating improvements in LLMs' ability to identify these salient dimensions across various model sizes. AI

IMPACT Enhances LLM capabilities in understanding and responding to emotional nuances in conversations.

RANK_REASON The cluster describes a new academic paper introducing a benchmark and a framework for LLMs. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• Bayesian inverse planning
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• Large Language Models
• PRISM
• ScienceCast

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →