DoNotAnswer
收藏资源简介:
该数据集包含文本内容以及对应的标签,用于区分内容是有益(Benign)还是有害(Harmful)。数据集分为三个配置:默认配置、neg配置和pos配置。每个配置都包括clf_label(分类标签)、instructions(指示说明)、content(文本内容)、answer_prompt(回答提示)、proxy_clf_label(代理分类标签)、gen_target(生成目标)和proxy_gen_target(代理生成目标)等字段。默认配置和pos配置的训练集包含132个示例,而neg配置的训练集为空。验证集在所有配置中都是空的。数据集可用于训练模型以进行文本分类任务。
This dataset contains text content and their corresponding labels for distinguishing whether the content is benign or harmful. The dataset is divided into three configurations: default configuration, neg configuration, and pos configuration. Each configuration includes fields such as clf_label (classification label), instructions, content (text content), answer_prompt, proxy_clf_label, gen_target, and proxy_gen_target. The training sets of the default configuration and pos configuration contain 132 examples, while the training set of the neg configuration is empty. The validation set is empty across all configurations. This dataset can be used for training models to perform text classification tasks.
数据集概述
基本信息
- 数据集名称: DoNotAnswer
- 存储库地址: https://huggingface.co/datasets/AlignmentResearch/DoNotAnswer
数据集配置
数据集包含以下三种配置:
1. default配置
- 特征:
clf_label: 分类标签(Benign或Harmful)instructions: 字符串类型content: 字符串序列answer_prompt: 字符串类型proxy_clf_label: int64类型gen_target: 字符串类型proxy_gen_target: 字符串类型
- 数据分割:
train: 132个样本,20918字节validation: 0个样本,0字节
- 下载大小: 9692字节
- 数据集大小: 20918字节
2. neg配置
- 特征: 同default配置
- 数据分割:
train: 0个样本,0字节validation: 0个样本,0字节
- 下载大小: 4268字节
- 数据集大小: 0字节
3. pos配置
- 特征: 同default配置
- 数据分割:
train: 132个样本,20918字节validation: 0个样本,0字节
- 下载大小: 9692字节
- 数据集大小: 20918字节
数据文件路径
- default配置:
train: data/train-*validation: data/validation-*
- neg配置:
train: neg/train-*validation: neg/validation-*
- pos配置:
train: pos/train-*validation: pos/validation-*




