Ayush-Singh/dberta-clusters

该数据集包含多个特征字段，如chosen（选择的选项）、rejected（被拒绝的选项）、reward_chosen（选择选项的奖励）、reward_rejected（被拒绝选项的奖励）、confidence（置信度）和cluster（聚类）。数据集被分为五个分片（cluster_0到cluster_4），每个分片都有相应的字节大小和示例数量。数据集的下载大小为171971694字节，总大小为314360171字节。

The dataset contains multiple feature fields such as chosen (selected option), rejected (rejected option), reward_chosen (reward for the chosen option), reward_rejected (reward for the rejected option), confidence (confidence level), and cluster (cluster). The dataset is divided into five splits (cluster_0 to cluster_4), each with corresponding byte sizes and example counts. The download size of the dataset is 171971694 bytes, and the total size is 314360171 bytes.