SKYLENAGE-ReasoningMATH

覆盖从基础算术到高等数学的多层次推理任务，涵盖逻辑推导、代数变换、几何分析、概率统计等多个维度，旨在评估模型在结构化数学问题中的理解、推理与解答能力。数据集设计注重问题难度分级与认知层次划分，支持细粒度的能力评估。

This dataset encompasses multi-level reasoning tasks spanning from basic arithmetic to advanced mathematics, covering multiple dimensions such as logical deduction, algebraic transformation, geometric analysis, probability and statistics. It aims to evaluate a model's capabilities in understanding, reasoning and solving structured mathematical problems. The dataset design prioritizes problem difficulty grading and cognitive hierarchy partitioning, supporting fine-grained capability assessment.