OnomaAIResearch/s1k_Korean

S1K-KR数据集是simplescaling/s1K数据集的韩语翻译，包含了1000个问题解决方案对，覆盖了多个领域。每个实例都包含了问题陈述、解决方案和详细的思维轨迹（步骤推理）。该数据集用于训练和评估韩语语言模型在复杂推理任务上的性能，支持的任务包括复杂推理、步骤推理提示、问题解决能力评估、韩语理解和生成、逐步解决方案生成等。

The S1K-KR dataset is a Korean translation of the simplescaling/s1K dataset, containing 1,000 problem-solution pairs across various domains. Each instance includes a problem statement, a solution, and detailed thinking trajectories (step-by-step reasoning). This dataset is designed for training and evaluating Korean language models on complex reasoning tasks, supporting tasks such as complex reasoning, step-by-step reasoning prompting, problem-solving capability evaluation, Korean language understanding and generation, and step-by-step solution generation.