Explore high-quality datasets for your AI and machine learning projects.
CBT‑Bench is a benchmark dataset aimed at evaluating large language models (LLMs) in assisting cognitive behavioral therapy (CBT). The dataset consists of three levels, each focusing on different key aspects of CBT, including basic knowledge recall, cognitive model understanding, and therapeutic response generation. Its goal is to assess LLMs’ support capability across various stages of professional mental health care, particularly CBT. The dataset includes multiple tasks and data files such as multiple‑choice questions, cognitive distortion classification, and therapeutic response generation exercises.
