kothasuhas/llama-3b-gold-15M-student-generations_SNIS

这是一个包含文本和数值特征的数据集，总共包含约14999750个训练样本和1000个验证样本。数据集的特征包括文本内容（text）和日志权重（log_weight）。文本内容是字符串类型，日志权重是32位浮点数类型。数据集的总大小约为35798GB，下载大小约为51661GB。This dataset includes text and numerical features, with a total of approximately 14,999,750 training samples and 1,000 validation samples. The features of the dataset include text content (text) and log weight (log_weight). The text content is of string type, and the log weight is of 32-bit floating-point type. The total size of the dataset is about 35,798GB, and the download size is about 51,661GB.