scottgeng00/olmo-3-preference-mix-deltas_reasoning_nothink-yolo_scottmix-DECON

该数据集包含多个字段，如提示语句(prompt)、选中的内容(chosen)和被拒绝的内容(rejected)，其中每个内容字段又包括内容和角色两个子字段。此外，还有选中和拒绝内容的模型名称、数据集名称、提示语句ID和类别等字段。数据集分为训练集，包含约294135个示例，大小约为2.26GB。The dataset includes multiple fields such as prompt, chosen, and rejected content. Each content field consists of sub-fields for content and role. Additionally, there are fields for the model names for chosen and rejected content, dataset name, prompt ID, and category. The dataset is split into a training set, which contains approximately 294135 examples and is about 2.26GB in size.