DukeCEICenter/Safety_Reasoning_Multi_Turn_Dialogue

该数据集包含了多轮对话中与安全推理相关的信息。具体特征包括全局ID、类别、目标、对话轮数、对话内容、cot（可能指代对话中的某些上下文或操作）、输出、问题标签和问题严重性等。数据集分为训练集，并提供相关配置文件。数据集的具体内容和用途没有在README中明确说明，但可以根据相关论文和引用进行推断。

The dataset contains information related to safety reasoning in multi-turn dialogues. Specific features include global ID, category, goal, turn number, conversation content, cot (which might refer to some context or operation in the dialogue), output, question label, and question severity, etc. The dataset is split into a training set and includes related configuration files. The specific content and purpose of the dataset are not explicitly described in the README, but can be inferred from the associated paper and citation.