HeartLink is an empathetic psychological model that uses a large language model fine-tuned on a large empathetic Q&A dataset. It can perceive users' emotions and experiences during conversations and provide empathetic responses using rich psychological knowledge, aiming to understand, comfort, and support users. The responses include emoji expressions to bridge the gap with users, offering psychological support and help during consultations.
The HeartLink project is based on the InternLM2-Chat model and has been fine-tuned to achieve empathetic functionality. The project supports text-to-speech synthesis and digital human display, and provides user emotion chart analysis. The project is under continuous development, and contributions through Star, PR, and Issue are welcome. The HeartLink psychological empathy question-and-answer dataset is derived from real psychological counseling scenarios. The first version uses about 180k rounds of question-and-answer pairs. The data covers a wide range of scenarios, including love, marriage, workplace, life, society, learning, sex, past, emotions, education, counseling, crisis, and many other rich scenarios.
SoulChat2.0 is a framework for constructing the digital twin of psychological counselors, designed to support the development of AI applications in mental health. It includes a data generation module and a modeling module, enabling the creation of personalized counseling models based on limited real-world counseling cases.
HappyDB is a crowd-sourced collection of 100,000 happy moments designed to advance the understanding of happiness through text analysis. The database is publicly available and aims to support research in natural language processing (NLP) and positive psychology. It provides insights into the causes of happiness and suggests sustainable actions for improving well-being.
Psy-Insight is a bilingual, interpretable multi-turn dataset for mental health counseling dialogues. It includes 6,208 rounds of multi-turn counseling dialogues in English and 5,776 rounds in Chinese, annotated with step-by-step reasoning labels and multi-task labels. This dataset is designed to support the application of large language models in mental health and is suitable for tasks such as emotion classification and psychological treatment interpretation.