symoon11/countdown-rl

该数据集包含了数据来源、提示（包括内容和角色）、能力、奖励模型（包括 ground_truth 和 style）、额外信息（包括索引、数量、解决方案、分割和目标）等字段。数据集被划分为训练集、验证集、测试集seen和测试集unseen。训练集包含200,000个示例，验证集包含1,000个示例，每个测试集也包含10,000个示例。数据集的下载大小为38MB，总大小为171MB。

The dataset includes fields such as data source, prompt (including content and role), ability, reward model (including ground_truth and style), and extra information (including index, numbers, solution, split, and target). The dataset is split into training set, validation set, test set seen, and test set unseen. The training set contains 200,000 examples, the validation set contains 1,000 examples, and each test set also contains 10,000 examples. The download size of the dataset is 38MB, and the total size is 171MB.