Menlo/Maze-Reasoning-Reset-v0.1

该数据集包含两个配置：order-1和order-2。每个配置都包括提示（Prompt）、上下文相关响应（Cot_Response）、实际响应（Response）、原始标记（raw_token）和对话信息（conversations，包含对话内容和角色）。order-1配置的训练集包含50705个示例，文件大小为1,372,910,725字节。order-2配置的训练集包含419,651个示例，文件大小为9,995,695,211字节。

The dataset consists of two configurations: order-1 and order-2. Each configuration includes a prompt, context-related response, actual response, raw tokens, and conversation information (including the content and role of the conversation). The training set for the order-1 configuration contains 50,705 examples and is 1,372,910,725 bytes in size. The training set for the order-2 configuration contains 419,651 examples and is 9,995,695,211 bytes in size.