🧠 OpenR1-Psy Dataset

This dataset is introduced in the paper: arXiv:2505.15715. For further details, please refer to our GitHub repository at Github. If you find this Dataset helpful, feel free to ⭐ it! OpenR1-Psy.

📘 Overview

OpenR1-Psy is a large-scale psychological counseling dataset that integrates diagnostic reasoning and therapeutic reasoning to train and evaluate large language models for mental health dialogue generation.
It goes beyond empathy-focused corpora by incorporating explicit reasoning traces grounded in DSM/ICD diagnostic standards and diverse psychotherapy frameworks such as CBT, ACT, psychodynamic, and humanistic therapy.

---

🔍 Key Features

• 🗂️ 19,302 dialogues with multi-turn psychological counseling interactions
• 🧩 Joint reasoning–response pairs: includes diagnostic & therapeutic reasoning traces
• 🧠 Clinically grounded: aligned with DSM/ICD standards
• 💬 Therapeutic diversity: CBT, ACT, psychodynamic, humanistic, and integrative approaches
• ✅ Quality-controlled: multi-dimensional validation for coherence, reasoning completeness, and framework consistency

---

🎯 Use Cases

• Fine-tuning LLMs for mental health counseling or psychological dialogue reasoning
• Evaluating models on clinical reasoning, empathy, and safety
• Building interpretable and evidence-based AI counselors (e.g., PsyLLM)

---

📄 Citation

If you use this dataset, please cite:

@article{hu2025beyond,
  title={Beyond Empathy: Integrating Diagnostic and Therapeutic Reasoning with Large Language Models for Mental Health Counseling},
  author={Hu, He and Zhou, Yucheng and Si, Juzheng and Wang, Qianning and Zhang, Hengheng and Ren, Fuji and Ma, Fei and Cui, Laizhong},
  journal={arXiv preprint arXiv:2505.15715},
  year={2025}
}

---

🧩 License

For research and educational use only.

Please ensure compliance with ethical and legal standards in mental health AI research.