scottgeng00/olmo-3-preference-mix-deltas_reasoning_nothink-yolo_victoria_hates_code

该数据集包含了一个提示（prompt）和两个选择（chosen与rejected），每个选择都包含了内容（content）和角色（role）。此外，数据集还记录了被选择和被拒绝的模型名称，数据来源的集名称，提示的唯一标识符以及类别。数据集分为训练集，其大小为2247269527.68字节，共有303757个示例。The dataset includes a prompt and two choices (chosen and rejected), each containing content and role. Additionally, the dataset records the names of the chosen and rejected models, the name of the dataset source, a unique identifier for the prompt, and the category. The dataset is split into a training set, which is 2247269527.68 bytes in size and contains 303757 examples.