RabidUmarell/concatenated_toxic_datasets
收藏Hugging Face2025-03-05 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/RabidUmarell/concatenated_toxic_datasets
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含文本及其相关负面行为标签的数据集，共有七个标签：文本内容(text)，身份攻击(identity_attack)，侮辱(insult)，粗俗(obscene)，威胁(threat)，毒性(toxicity)和严重毒性(severe_toxicity)。每个标签对应一个浮点数类型的数据，表示该标签在文本中的程度。数据集分为训练集，包含848862个样本。
This dataset consists of text and its associated labels for negative behaviors, with a total of seven labels: text content (text), identity attack (identity_attack), insult (insult), obscenity (obscene), threat (threat), toxicity (toxicity), and severe toxicity (severe_toxicity). Each label corresponds to a floating-point number indicating the degree of that label in the text. The dataset is split into a training set with 848862 samples.
提供机构：
RabidUmarell
