[2310.07146] Empowering Psychotherapy with Large Language Models: Cognitive Distortion Detection through Diagnosis of Thought Prompting

Cognitive Behavior Therapy (CBT) is an important form of psychotherapy that involves discovering the patient’s maladaptive cognitive process and providing guidance to a more balanced thinking. One of the early steps in CBT is to identify the underlying cognitive distortion, such as all-or-none thinking. This is based on the patient’s conversation with the therapist, which can possibly be assisted by an LLM.

The authors propose Diagnosis of Thought (DoT) prompting and show its performance gain in detecting the underlying cognitive distortion.

The Diagnosis of Thought (DoT) Prompting

DoT is inspired by actual therapists’ process in detecting the cognitive distortion, which follows 1) subjectivity assessment, 2) contrastive reasoning, and 3) schema analysis. The experimental results indicate that GPT4+DoT achieves the best performance.

• Knowledge-Augmented Language Model Verification
• Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
• Video Language Planning
• PaLI-3 Vision Language Models: Smaller, Faster, Stronger
• Large Language Models Are Zero-Shot Time Series Forecasters