akcit-ijf/jigsaw-toxic-comment-train-processed-seqlen128_translated

该数据集包含多条评论，每条评论具有唯一的id和评论内容（comment_text），并且标注了评论是否具有毒性（toxic）、严重毒性（severe_toxic）、淫秽（obscene）、威胁（threat）、侮辱（insult）、身份仇恨（identity_hate）等属性。此外，还包含输入词id（input_word_ids）、输入掩码（input_mask）和所有段id（all_segment_id）等特征。数据集分为训练集，包含223,549个样本，总大小为423,101,877字节。

This dataset contains multiple features, primarily used for analyzing and classifying undesirable attributes in comment text. The features include a unique identifier for each comment (id), the comment text itself (comment_text), and several labels (toxic, severe_toxic, obscene, threat, insult, identity_hate) that indicate whether the comment contains the respective undesirable attributes. Additionally, the dataset includes extra features for model input, such as input_word_ids, input_mask, and all_segment_id. The dataset is divided into a training set, containing 223549 samples.