mlfoundations-dev/pos_neg_ablation_instruction_filtering_seed_code_best_ioi

该数据集包含了指令种子(instruction_seed)、来源(source)、推理过程(reasoning)、DeepSeek解决方案(deepseek_solution)、原始行索引(__original_row_idx)、最终推理轨迹(final_reasoning_trace)和对话(conversations)等字段。这些字段表明数据集可能与对话系统或推理任务相关。训练集包含10000个示例，数据集总大小为707MB。具体的数据集用途和详细内容描述未在README中提供。The dataset includes fields such as instruction_seed, source, reasoning, deepseek_solution, __original_row_idx, final_reasoning_trace, and conversations. These fields suggest that the dataset may be related to dialogue systems or reasoning tasks. The training set contains 10,000 examples, and the total size of the dataset is 707MB. The specific purpose and detailed content description of the dataset are not provided in the README.