mlfoundations-dev/r1_rewriting_math_with_gt

该数据集包含问题（problem）、推理过程（reasoning）、deepseek解决方案（deepseek_solution）、真实解决方案（ground_truth_solution）、是否正确（correct）、评判推理（judge_reasoning）、新推理（new_reasoning）、原始行索引（__original_row_idx）、对话（conversations，包含发送者和消息内容）、系统提示（system_prompt）等字段。训练集（train）包含5个示例，数据集大小为199716字节。

The dataset includes fields such as problem, reasoning, deepseek solution, ground truth solution, correctness (correct), judgment of reasoning (judge_reasoning), new reasoning (new_reasoning), original row index (__original_row_idx), conversation (including sender and message content), and system prompt (system_prompt). The training set (train) contains 5 examples, and the total size of the dataset is 199716 bytes.