Cognitive Reframing via Large Language Models for Enhanced Linguistic Attributes

Xiaomeng Wang, Dharmendra Sharma, Dinesh Kumar

Keywords: Cognitive Reframing, LLMs, Language Agent

TL;DR: Leveraging Large Language Models to enhance the quality of Cognitive Reframing.

Abstract: Cognitive Reframing, a core technique in Cognitive Behavioral Therapy (CBT), seeks to enhance mental well-being. While previous research has highlighted the efficacy of Large Language Models (LLMs) for cognitive reframing, there has been limited focus on enhancing reframing quality across multiple linguistic attributes in the final output. This paper fills this gap by employing LLMs to generate and iteratively refine reframed thoughts. The results of our study outperform in helpfulness, empathy and rationality in GPT-4 evaluation.

Submission Number: 107