PrimeIntellect/Reverse-Text-SFT

这是一个小型SFT数据集，用于预训练小型模型，如Qwen/Qwen3-0.6B。数据集包含了按照prompt-completion格式组织的示例，每个示例是将5-20个单词的文本字符顺序颠倒。该数据集由willcb/R1-reverse-wikipedia-paragraphs-v1-1000数据集的文本处理而成。This is a small SFT dataset for pre-training small models, such as Qwen/Qwen3-0.6B. The dataset contains examples in prompt-completion format, each example is text with 5-20 words reversed character-by-character. The dataset is processed from the text of willcb/R1-reverse-wikipedia-paragraphs-v1-1000.