HR-MultiWOZ
收藏资源简介:
HR-MultiWOZ是由亚马逊创建的,专为HR领域的大型语言模型(LLM)代理训练和评估设计的对话数据集。该数据集包含550个完全标注的对话,覆盖10个HR相关领域,旨在解决HR领域中对话数据稀缺和隐私问题。数据集的创建过程主要依赖于LLM,辅以最小程度的人工标注,确保了时间和成本效率。HR-MultiWOZ不仅为HR自动化提供了丰富的、富有同情心的对话数据,还为其他领域的标注对话数据生成提供了一个可转移的方法。
HR-MultiWOZ is a conversational dataset developed by Amazon, specifically designed for training and evaluating large language model (LLM) agents in the human resources (HR) domain. This dataset contains 550 fully annotated conversations spanning 10 HR-related fields, aiming to address the scarcity of conversational data and privacy concerns within the HR sector. The creation of HR-MultiWOZ primarily relies on LLMs, supplemented with minimal manual annotation, which ensures time and cost efficiency. HR-MultiWOZ not only provides rich, empathetic conversational data for HR automation but also offers a transferable methodology for generating annotated conversational data across other domains.




