HKUST-DSAIL/Graph-R1-RFT-COT-30K

Graph-CoT-30k是一个大规模、高质量的教学调整数据集，旨在增强大型语言模型在复杂图论问题上的推理能力。该数据集包含30000个问题-答案对，每个对都包含由QwQ-32B模型生成的超长链式思维推理轨迹。它旨在通过高质量的思维过程数据，教授模型如何像专家一样对具有挑战性的组合优化和图结构分析任务进行详细、多步骤、逻辑严谨的推理。

Graph-CoT-30k is a large-scale, high-quality instruction tuning dataset designed to enhance the reasoning capabilities of large language models (LLMs) on complex graph-theoretic problems. It contains 30,000 question-answer (QA) pairs, each featuring ultra-long chain-of-thought (CoT) reasoning traces generated by the state-of-the-art reasoning model QwQ-32B. The dataset aims to teach models to reason like experts by producing detailed, multi-step, logically rigorous solutions to challenging combinatorial optimization and graph structure analysis tasks through high-quality thought process data.