ansulev/DeepSeek-V4-Distill-8000x
收藏资源简介:
DeepSeek-V4-Distill-8100x是一个用于推理导向蒸馏的监督微调数据集。问题提示来源于[Jackrong/GLM-5.1-Reasoning-1M-Cleaned](https://huggingface.co/datasets/Jackrong/GLM-5.1-Reasoning-1M-Cleaned),答案由教师模型DeepSeek-V4-Flash生成。经过清洗后,发布的`train`分割包含7,716个高质量的JSONL示例。数据集格式包括对话式和直接输入/输出字段,主要用于推理导向的监督微调、蒸馏实验和格式转换实验。数据集存在一些局限性,如可能包含教师模型的错误或偏见。
DeepSeek-V4-Distill-8100x is a supervised fine-tuning dataset for reasoning-oriented distillation. The question prompts come from [Jackrong/GLM-5.1-Reasoning-1M-Cleaned](https://huggingface.co/datasets/Jackrong/GLM-5.1-Reasoning-1M-Cleaned), and the answers were generated by the teacher model DeepSeek-V4-Flash. After the cleaning process, the released `train` split contains 7,716 high-quality JSONL examples. The dataset format includes both conversation-style and direct input/output fields, and it is primarily intended for reasoning-oriented supervised fine-tuning, distillation experiments, and format conversion experiments. The dataset has some limitations, such as potential factual errors or biases inherited from the teacher model.




