juliadollis/Qwen3-8B_10ep_ok_2_test_predictions

该数据集包含与总统访谈相关的记录，涵盖标题、日期、总统姓名、URL、问题顺序、访谈问题、访谈回答等多个字段。还包括GPT-3.5生成的摘要和预测、问题、注释者信息、不可听标记、多问题标记、肯定问题标记等。数据集还包含清晰度和回避标签，以及标准化和预测的清晰度标签。数据集仅包含训练集，共308个示例。This dataset contains records related to presidential interviews, including fields such as title, date, presidents name, URL, question order, interview question, interview answer, and more. It also includes GPT-3.5 generated summaries and predictions, questions, annotator information, inaudible markers, multiple questions markers, affirmative questions markers, etc. The dataset also includes clarity and evasion labels, as well as normalized and predicted clarity labels. The dataset contains only a training set with 308 examples.