KOREAson/YiSang-STEM_Code-Unfiltered

YiSang-STEM_Code-Unfiltered是一个包含1.1M条长文本推理轨迹的数据集，这些轨迹是通过Qwen3-32B模型生成的。数据集包含128,524个与STEM或编码主题相关的独特韩文提示，是从网络上收集的。这是一个更大且未经过滤的数据集版本，可能会用于未来的迭代。YiSang-STEM_Code-Unfiltered is a collection of 1.1M long-cot reasoning traces generated via Qwen3-32B. It consists of 128,524 unique Korean prompts related to STEM or Coding topics collected from the web. This is a bigger and unfiltered version of the dataset, which might be used in our future iterations.