theprint/SelfHelp-ShareGPT
收藏资源简介:
SelfHelp是一个通过DataMix生成的精选对话数据集,旨在支持心理健康教育和心理自助技术领域的语言模型训练与评估。该数据集包含多轮对话,涵盖认知行为疗法(CBT)、识别自动负面思维、正念练习以及为学生和从业者设计治疗工作坊活动等主题。每个条目都包含结构化、基于教学原理的响应,模拟循证推理、逐步指导和反思性提问,适用于教育聊天机器人、治疗师培训工具或心理健康素养项目等应用。该数据集与现实世界的治疗框架和教学设计原则紧密对齐,对于构建能够准确呈现临床心理学概念并保持非专业用户可访问性的AI系统的研究人员和开发者特别有价值。
SelfHelp is a curated collection of conversational data generated by DataMix, designed to support training and evaluation of language models in the domain of mental health education and psychological self-help techniques. The dataset features multi-turn dialogues covering topics such as cognitive behavioral therapy (CBT), identifying automatic negative thoughts, mindfulness practices, and designing therapeutic workshop activities for students and practitioners. Each entry includes structured, pedagogically grounded responses that model evidence-based reasoning, step-by-step instruction, and reflective questioning—ideal for applications in educational chatbots, therapist training tools, or mental health literacy programs. With clear alignment to real-world therapeutic frameworks and instructional design principles, this dataset is particularly valuable for researchers and developers building AI systems that must accurately represent clinical psychology concepts while maintaining accessibility for non-expert users.




