MathScale: Scaling Instruction Tuning for Mathematical Reasoning

arXiv:2403.02884v1 Announce Type: cross
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities in problem-solving. However, their proficiency in solving mathematical problems remains inadequate. We propose MathScale, a simple and scalable method to create high-quality mathematical reasoning data using frontier LLMs (e.g., {\tt GPT-3.5}). Inspired by the cognitive mechanism in human mathematical learning, it first extracts topics and knowledge points from seed math questions and then build a concept graph, which is subsequently used to generate new math questions. MathScale …

abstract arxiv capabilities cognitive cs.ai cs.cl cs.lg data gpt gpt-3 gpt-3.5 human language language models large language large language models llms mathematical reasoning problem-solving quality reasoning scalable scaling simple type