zjhhhh/iter2_multi_scores_adversary_rescaled

该数据集包含了一系列对话相关的字符串字段和用于分析对话选择的浮点数字列表。具体包括提示信息(prompt)、需求(requirements)、不同选择的响应(selection_response)以及当前和对手的响应(current_response和adversary_response)。此外，还有基础响应(base_response)和多个表示不同选择与对手之间平均分和多数情况的字段。数据集分为训练集(train)，共有约50057个示例。

The dataset consists of a series of dialogue-related string fields and float list fields for analyzing dialogue selections. It includes prompt information, requirements, selection responses (selection_response), current and adversary responses (current_response and adversary_response). Additionally, there are base responses (base_response) and multiple fields representing the average scores and majority situations between different selections and adversaries. The dataset is split into a training set (train) with approximately 50,057 examples.