BianQue

Why we included this project

BianQue tackles a specific problem most general chatbots get wrong: gathering a useful medical history. It was fine-tuned on tens of millions of Chinese health dialogue samples with an emphasis on the asking side of a consultation, so it follows up with clarifying questions rather than jumping to an answer. That makes it a sensible base for researchers prototyping Chinese-language health chatbots, and you can check its behavior quickly since the weights are on Hugging Face alongside a live demo. If mental health is part of the picture, the companion SoulChat model covers empathetic, multi-turn counseling, so the two give you complementary domain-tuned starting points. Treat it as a 2023 research release rather than a maintained product; use it to inform experiments, not to replace a clinician's judgment.