parth51/dealignment-dataset

该数据集是一个文本分类任务的数据集，主要包含英文文本。数据集的大小在10K到100K之间，具体包含14,000个训练样本和5,600个测试样本。数据集的特征包括accepted_text、rejected_text和prediction，均为字符串类型。数据集是从annotate-anthropic-hf数据集中提取的一个类别平衡的子集。

This dataset is a text classification dataset primarily containing English text. The dataset size is between 10K and 100K, specifically including 14,000 training samples and 5,600 test samples. The dataset features include accepted_text, rejected_text, and prediction, all of which are of string type. The dataset is a class-balanced subset extracted from the annotate-anthropic-hf dataset.