mlfoundations-dev/difficulty_sorting_easy_seed_math

该数据集包含了指令种子(instruction_seed)、来源(source)、难度(difficulty)、难度理由(difficulty_reasoning)、原始行索引(__original_row_idx)、模型响应(r1_distill_70b_response)以及对话(conversations)等信息。它被划分为训练集(train)，具有1000个示例，文件大小为19518548字节。但是具体的应用场景和详细描述在README中并未给出。

The dataset includes fields such as instruction_seed, source, difficulty, difficulty_reasoning, __original_row_idx, model response (r1_distill_70b_response), and conversation (conversations). It is split into a training set (train) with 1000 examples and a file size of 19518548 bytes. However, the specific application scenario and detailed description are not provided in the README.