zjhhhh/iter2_scores_adversary_rescaled

该数据集包含了多个文本字段，如提示(prompt)、要求(requirements)和多个选择响应(selection_response)，以及当前响应(current_response)和对手响应(adversary_response)。此外，还包含了基础响应(base_response)和不同选择的对手平均(mean)及多数(majority)统计数据。数据集仅包含训练集分割，共有50055个示例。The dataset includes multiple text fields such as prompt, requirements, multiple selection responses (selection_response), current responses (current_response), and adversary responses (adversary_response). It also contains base responses (base_response) and statistical data for the mean and majority of different selections against adversaries. The dataset consists only of a training split with a total of 50,055 examples.