INSPIRED2
收藏资源简介:
INSPIRED2是由克拉根福大学创建的数据集,专注于社交对话推荐,包含1001个对话,平均每对话10.73轮,平均每轮7.93个tokens。数据集通过手动修正INSPIRED数据集中的标注错误而创建,旨在提高对话推荐系统的性能。INSPIRED2数据集的应用领域主要集中在提升对话推荐系统的质量和用户交互的自然性,通过改进数据质量来增强系统的响应准确性和一致性。
INSPIRED2 is a dataset developed by the University of Klagenfurt, focusing on social conversational recommendation. It contains 1001 conversations, with an average of 10.73 turns per conversation and 7.93 tokens per turn on average. This dataset was constructed by manually correcting annotation errors in the original INSPIRED dataset, aiming to improve the performance of conversational recommendation systems. The core application scenarios of INSPIRED2 center on enhancing the quality of conversational recommendation systems and the naturalness of user interactions, as well as strengthening the response accuracy and consistency of such systems by improving data quality.




