mlfoundations-dev/embedding_apply_automath

该数据集包含了多个特征字段，如指令种子（instruction_seed）、来源（source）、嵌入表示（embeddings）、正面评分平均值（mean_positive_score）、负面评分平均值（mean_negative_score）和分数差（difference_score）。数据集分为训练集（train），包含4000个示例，大小为34740030.4字节。整个数据集的下载大小为27273939字节。但是，数据集的具体内容和用途没有在README中描述，因此无法提供详细的数据集中文描述。

The dataset contains multiple feature fields such as instruction_seed, source, embeddings, mean_positive_score, mean_negative_score, and difference_score. The dataset is split into a training set (train) with 4000 examples, totaling 34740030.4 bytes in size. The entire dataset has a download size of 27273939 bytes. However, the specific content and purpose of the dataset are not described in the README, so no detailed dataset description can be provided.