Nidhogg-zh/Multi-Questioner_Dialogue

Multi-Questioner Dialogue (MQDialog)数据集旨在促进问题者感知个性化研究。该数据集包含来自英语和中文电视剧剧本及现实世界对话的多轮对话，其中选定主要演员作为回答者，其他角色或联系人作为提问者。数据集包含12个回答者和173个提问者，支持对话生成、响应评估和多问题者设置下的对话模型性能基准测试。数据集来源包括《生活大爆炸》、《老友记》、《摩登家庭》等英语剧本，以及《武林外传》、《甄嬛传》等中文剧本，还包括来自微信的两人聊天记录。数据集结构包括回答者、提问者以及训练和测试集的划分。

The Multi-Questioner Dialogue (MQDialog) dataset is designed to facilitate research in questioner-aware personalization. It contains dialogues with various questioners for each responder. The dataset is derived from English and Chinese scripts of popular TV shows and real-world conversations. It includes dialogues where selected leading actors act as responders, while other characters or contacts serve as questioners. The dataset contains a diverse set of 12 responders and 173 questioners. The dataset supports research on dialogue generation, response evaluation, and questioner-aware personalization in multi-turn conversations.