leyond123/test_prefenence_datasets

该数据集包含文本生成相关的指令、话题和生成文本，以及针对生成文本的评分和评分理由。数据集还包含了模型名称和用于 ultra feedback 的原始输入、输出及统计信息。数据集分为两个配置：format_text_generation_d_p_o_0 和 preference_to_argilla_0，每个配置下都有训练集数据。

The dataset includes text generation related instructions, topics, and generated texts, as well as ratings and rationales for the generated texts. The dataset also contains model names and raw input, output, and statistical information for ultra feedback. The dataset is divided into two configurations: format_text_generation_d_p_o_0 and preference_to_argilla_0, each with training set data.