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

该数据集包含对话提示(prompt)、选中的回答(chosen)、被拒绝的回答(rejected)、回答的模型(chosen_model和rejected_model)、数据集名称(dataset)、提示ID(prompt_id)和类别(category)等信息。chosen和rejected字段分别包含了回答内容和回答角色。数据集分为训练集(train)，共有278,757个示例，大小为6.19GB。The dataset includes conversation prompts (prompt), chosen responses (chosen), rejected responses (rejected), response models (chosen_model and rejected_model), dataset name (dataset), prompt ID (prompt_id), and category (category). The chosen and rejected fields each contain response content and response roles. The dataset is split into a training set (train) with a total of 278,757 examples and a size of 6.19GB.