Charismatic Leadership Tactics Dataset

该数据集包含了大约1.2万个样本，均匀分布在十二种魅力领导策略（CLTs）中，这些样本是通过ChatGPT API生成的。数据集涵盖了多样化的提示和场景，以确保各种魅力策略能够得到均衡的体现。在整理数据集时，我们严格确保样本符合目标魅力领导策略，并应用了严格的质量控制流程，以保证生成的文本样本在准确性和一致性上的高标准。该数据集大约包含12万个句子（每种策略1万个），其任务是检测自然语言中的魅力领导策略。This dataset contains approximately 12,000 samples evenly distributed across 12 charismatic leadership tactics (CLTs), all generated via the ChatGPT API. To ensure balanced representation of each charismatic leadership tactic, the dataset covers diverse prompts and scenarios. During the dataset curation process, we strictly ensured that every sample matched its target charismatic leadership tactic, and applied rigorous quality control workflows to maintain high standards of accuracy and consistency for the generated textual samples. The dataset comprises roughly 120,000 sentences (10,000 sentences per tactic), and its core task is to detect charismatic leadership tactics in natural language.

搜集汇总

背景与挑战

背景概述

该数据集包含美国9位女性和9位男性州长在COVID-19危机前后各7篇演讲，共252篇，用于分析危机情境下魅力领导策略（CLT）的使用情况。数据集通过BERT模型对演讲文本进行标注，涵盖隐喻、道德信念、修辞问题等多种策略，旨在研究危机对领导者魅力的影响及性别差异。