mlfoundations-dev/instruction_filtering_random_seed_data_code_w_openthoughts

这是一个包含对话系统交互数据的训练集，其中包括系统信息，对话内容（包括发送者和消息内容），问题，推理过程，DeepSeek解决方案，唯一标识符，真实解决方案，来源，代码（可能为空），正确性，评价推理，原始行索引和指令种子等信息。数据集大小为3.1GB，共有120,668个示例。

This is a training set containing interaction data of a dialogue system, including system information, conversation content (including sender and message content), problem, reasoning process, DeepSeek solution, unique identifier, ground truth solution, source, code (possibly null), correctness, reasoning evaluation, original row index, and instruction seed. The dataset size is 3.1GB with a total of 120,668 examples.