txchmechanicus/GLM-5.1-Reasoning-1M-Cleaned
收藏资源简介:
GLM-5.1-Reasoning-1M-Cleaned 是一个经过清理和重新格式化的数据集,源自 GLM-5.1-1000000x。它保留了原始数据集的四个子集(main、PHD-Science、Multilingual-STEM、Math),并将每个示例转换为统一的 SFT-ready 模式,包含明确的 `conversations`、`input`、`output`、`domain` 和 `meta` 字段。该数据集主要用于文本生成和问答任务,特别关注推理、思维链、指令调优和蒸馏等技术。清理过程移除了不完整、重复或无法解析的记录,并提供了详细的统计信息。
GLM-5.1-Reasoning-1M-Cleaned is a cleaned and reformatted derivative of the GLM-5.1-1000000x dataset. It preserves the original four-subset layout (main, PHD-Science, Multilingual-STEM, Math) while converting every example into a unified SFT-ready schema with explicit `conversations`, `input`, `output`, `domain`, and `meta` fields. The dataset is designed for text-generation and question-answering tasks, with a focus on reasoning, chain-of-thought, instruction-tuning, and distillation. The cleaning process removed incomplete, duplicated, or unparseable records, and detailed statistics are provided for each subset.




