RoleplayPref

该数据集名为RoleplayPref，是一个角色扮演对话偏好数据集，包含了1,108个角色和16,888段双语对话。其中训练集包含16,388段对话，测试集包含500段，角色涵盖了13个类别。该数据集的任务是为角色扮演对话代理进行偏好建模，规模达到了16,888段对话和1,108个角色。

RoleplayPref is a roleplay dialogue preference dataset. It encompasses 1,108 characters and 16,888 bilingual dialogue instances. The dataset is split into a training set with 16,388 dialogues and a test set with 500 dialogues, and its characters fall into 13 categories. The core task of this dataset is to perform preference modeling for roleplay dialogue agents, with a total scale of 16,888 dialogues and 1,108 characters.