mc-ai/counsel-chat-synthetic-conv-trees

这是一个由种子用户问题和助手角色构建的合成治疗对话数据集，每一轮对话中用户或助手可以选择5种对话策略之一进行回应。该数据集旨在作为人类偏好注释的源数据集，用于训练出在困难对话中产生同理心回应的奖励模型。

This is a synthetic therapy conversation dataset constructed using seed user questions and assistant personas, where the user or the assistant can choose one of the five conversational tactics to respond in each turn. The dataset is intended to serve as a source for human preference annotations to train a reward model for empathetic responses in challenging conversations.