scottgeng00/olmo-3-preference-mix-deltas_reasoning-chosen_qwen4b-yolo_scottmix-DECON

该数据集包含了文本提示(prompt)、选中的文本内容(chosen)和被拒绝的文本内容(rejected)，每个内容都包含了文本本身和角色信息。此外，还有选中文本和拒绝文本所使用的模型信息，数据集名称以及提示的唯一标识。数据集被划分为训练集(train)，共有293514个示例。由于README中没有提供详细的中文描述，这里根据字段内容进行简要描述。

The dataset includes text prompts, chosen content, and rejected content, each with its own text and role information. It also contains information about the models used for choosing and rejecting content, the name of the dataset, and a unique identifier for the prompt. The dataset is split into a training set (train) with a total of 293514 examples. Since the README does not provide a detailed description, this is a brief description based on the field contents.