selfcorrexp2/llama31_gold_baseline_tmp07

该数据集包含了索引（idx）、真实值（gt）、提示（prompt）、难度级别（level）、类型（type）、解决方案（solution）、个人解决方案（my_solu）、预测值（pred）和奖励（rewards）等字段。数据集被划分为训练集（train），共有15000个示例，文件大小为80462574字节。但具体的应用场景和详细描述未在README中提供。

The dataset includes fields such as index (idx), ground truth (gt), prompt, difficulty level (level), type, solution, personal solution (my_solu), prediction (pred), and rewards. The dataset is split into a training set (train) with 15,000 examples and a file size of 80,462,574 bytes. However, the specific application scenario and detailed description are not provided in the README.