Qulac

该数据集名为Qulac，是首个在信息检索环境中构建的澄清问题数据集，基于2009至2012年TREC网络跟踪数据而建立。它包含了超过1万个问题-答案对，覆盖了198个TREC主题和762个方面。该数据集的创建旨在通过主动提出澄清问题，提升对话系统中的用户满意度。其规模达到了为198个TREC主题提供超过1万对问题-答案。该数据集的任务是在开放领域的信息寻求对话中提出澄清问题。

This dataset, named Qulac, is the first clarifying question dataset constructed within the information retrieval environment, built based on the TREC Web Track data from 2009 to 2012. It contains over 10,000 question-answer pairs, covering 198 TREC topics and 762 aspects. This dataset was created to improve user satisfaction in conversational systems by proactively posing clarifying questions. Its scale amounts to over 10,000 question-answer pairs for 198 TREC topics. The task of this dataset is to generate clarifying questions in open-domain information-seeking dialogues.