MathFusionQA is a mathematical reasoning dataset that focuses on multi-step reasoning and solving of mathematical problems. The related paper results are:MathFusion: Enhancing Mathematical Problem-solving of LLM through Instruction Fusion".

The dataset contains 59,000 high-quality math question-answering samples, covering a variety of question types such as arithmetic operations, algebraic equations, geometric applications, logical reasoning, etc. Each sample contains a question (query), a detailed answer (response) and auxiliary indexes (idx_self, idx_neighbor). The question scenarios are rich, covering daily applications (such as shopping calculations, itinerary planning), academic training (such as geometric proofs, probability calculations), etc., aiming to improve the mathematical problem-solving ability of the large language model (LLM).