mlfoundations-dev/pos_neg_ablation_instruction_filtering_seed_math_best_openr1math

该数据集包含了多个特征字段，如指令种子(instruction_seed)、来源(source)、推理(reasoning)、解决方案(deepseek_solution)等。数据集分为训练集(train)，共有10000个示例。数据集的总大小为460410883字节，下载大小为198617591字节。

The dataset includes multiple feature fields such as instruction seed (instruction_seed), source (source), reasoning (reasoning), solution (deepseek_solution), etc. The dataset is split into a training set (train) with a total of 10,000 examples. The total size of the dataset is 460,410,883 bytes, and the download size is 198,617,591 bytes.