TAUR-dev/agg__BC__math500_r1_evals

该数据集包含四个浮点数特性：向后链接(backward_chaining)、验证(verification)、回溯(backtracking)和子目标(subgoals)。数据集目前只有一个训练集split，大小为32字节，包含1个数据示例。整个数据集的下载大小为2067字节。

The dataset includes four floating-point features: backward_chaining, verification, backtracking, and subgoals. The dataset currently has one train split, which is 32 bytes in size and contains 1 example. The total download size of the dataset is 2067 bytes.