LafouCC/civil-comments-merged

该数据集包含多个特征字段，如output、toxicity、severe_toxicity等，这些字段的数据类型包括字符串和浮点数。数据集分为一个训练集，包含2000599个样本，总大小为725558114字节。下载大小为427036778字节。数据集的配置信息显示，数据文件位于train分割路径下。

This dataset contains multiple feature fields such as output, toxicity, severe_toxicity, etc., with data types including string and float64. The dataset is divided into one training set containing 2000599 samples, with a total size of 725558114 bytes. The download size is 427036778 bytes. The configuration information of the dataset shows that the data files are located under the train split path.

原始信息汇总

数据集概述

数据特征

• output: 类型为字符串 (string)
• toxicity: 类型为浮点数 (float64)
• severe_toxicity: 类型为浮点数 (float64)
• obscene: 类型为浮点数 (float64)
• threat: 类型为浮点数 (float64)
• insult: 类型为浮点数 (float64)
• identity_attack: 类型为浮点数 (float64)
• sexual_explicit: 类型为浮点数 (float64)
• instruction: 类型为字符串 (string)

数据分割

• train: 包含2,000,599个样本，占用725,558,114字节

数据集大小

• 下载大小: 427,036,778字节
• 数据集大小: 725,558,114字节

配置

• config_name: default
  • data_files:
    • split: train
    • path: data/train-*