selfcorrexp2/llama31_sft_non_delete_300k

该数据集包含了多个字段，用于存储索引、提示、答案、对话轮次信息、对话内容、角色、正确答案、奖励、用户解决方案、标记、对话轮数、问题、上下文信息、难度级别、类型、预测结果和代码片段等。数据集分为训练集，共有300000个示例。具体应用场景和详细描述未在README中提供。

The dataset includes multiple fields for storing index, prompt, answers, conversation round information, conversation content, role, ground truth, rewards, user solutions, flags, turn number, questions, context information, difficulty level, type, prediction results, and code snippets. The dataset is split into a training set with a total of 300,000 examples. The specific application scenario and detailed description are not provided in the README.