DatPySci/weak_to_strong_reward_hh

这是一个包含文本和分数的数据集，具体包含的字段有：提示文本(prompt)、选中文本(chosen)、拒绝文本(rejected)、选中分数(chosen_score)、拒绝分数(rejected_score)、是否交换(is_swap)、上下文(ctx)和目标序列(target)。数据集分为训练集(train)、待排序集(w2s)和测试集(test)，每个集合的大小和示例数量都有详细说明。

This dataset contains text and score information, including fields such as prompt, chosen, rejected, chosen_score, rejected_score, is_swap, ctx, and target. The dataset is divided into three parts: training set (train), waiting-to-sort set (w2s), and test set (test), with detailed information about the size and number of examples in each set.