Xerv-AI/GRAD
收藏官方服务：
资源简介：
GRAD是一个高质量的合成数学数据集，包含1,933个原创问题，每个问题都附有详细的逐步证明。这些问题和证明都是从头开始生成的，没有出现在任何教科书、竞赛档案或研究论文中。数据集适用于数学专业语言模型的微调、长上下文推理和证明生成的评估、思维链系统的训练和基准测试，以及自动定理证明和数学推理的研究。
GRAD is a high-quality synthetic mathematics dataset containing 1,933 original problems, each accompanied by a complete, detailed, step-by-step proof. All problems and proofs were generated from scratch. No entry has ever appeared in textbooks, competition archives, or research papers. The dataset is intended for fine-tuning of mathematics-specialised language models, evaluation of long-context reasoning and proof generation, training and benchmarking chain-of-thought systems, and research in automated theorem proving and mathematical reasoning.
提供机构：
Xerv-AI
