APIGen-MT Synthetic Data

该数据集通过一个两阶段框架生成，捕捉了真实的人机交互动态，其中包含了合成的多轮交互数据。该数据集旨在通过提供高质量、可验证的交互，提升人工智能代理的训练效果。所涉及的任务是人工智能代理与人类用户之间的多轮交互生成。This dataset is constructed via a two-stage framework, which captures real-world human-computer interaction dynamics and includes synthesized multi-turn interaction data. It aims to enhance the training of AI Agents by providing high-quality, verifiable interaction samples. The core task covered by this dataset is the generation of multi-turn interactions between AI Agents and human users.