ZTE-AIM/32B_LLM_AdaptiveMath_data

该数据集名为32B_LLM_AdaptiveMath_data，用于训练LLM-Adaptive-ZMath-model-32B模型。数据集通过LLM-Adaptive Question Difficulty Grading方法构建，能够根据目标LLM的问题解决能力动态调整问题难度，以适应性地生成高质量的链式思维（CoT）数据。这些数据用于提升数学问题解决和代码生成等多个领域的推理性能。The dataset named 32B_LLM_AdaptiveMath_data is used for training the LLM-Adaptive-ZMath-model-32B. The dataset is constructed through the LLM-Adaptive Question Difficulty Grading method, which dynamically adjusts the difficulty of problems based on the target LLMs problem-solving capabilities to generate high-quality chain-of-thought (CoT) data. This data is used to enhance reasoning performance across multiple domains such as mathematical problem solving and code generation.