Debate Speeches Dataset

该数据集是一个包含3,685场辩论演讲的大型数据集，这些演讲都被标注了针对辩论关系的反驳内容。该数据集涵盖了针对不同议题的支持和反对演讲，这些演讲被分为明确反驳和隐含反驳两类。此外，数据集还包括了人工校正的转录文本和录音文件。其中，支持性演讲有1,797场，反对性演讲有1,888场，这些演讲中包含了明确和隐含的反驳内容。该数据集的规模为3,685场演讲，其任务目标是检测能够有效反驳给定文本中提出的论点的反驳演讲。

This is a large-scale dataset comprising 3,685 debate speeches, each annotated with rebuttal content aligned to their debate relationships. The dataset encompasses pro and con speeches across diverse topics, which are classified into two categories: explicit rebuttals and implicit rebuttals. Additionally, the dataset includes manually corrected transcriptions and corresponding audio recordings. Specifically, there are 1,797 pro speeches and 1,888 con speeches, both of which contain explicit and implicit rebuttal content. The total scale of this dataset is 3,685 speeches, and its core task objective is to detect rebuttal speeches that can effectively refute the arguments presented in a given text.