prithivMLmods/Poseidon-Reasoning-5M

Poseidon-Reasoning-5M是一个高质量、紧凑型的推理数据集，专为数学、编程和科学领域的高级应用而设计。它强调数学和一般推理挑战，适合用于大型语言模型研究、基准测试和STEM-focused教育工具。数据集包含了约500万个条目，采用箭头和Parquet格式，以英文为主。

Poseidon-Reasoning-5M is a high-quality, compact reasoning dataset designed for advanced applications in mathematics, coding, and science. It emphasizes mathematical and general reasoning challenges and is suitable for large language model research, benchmarking, and STEM-focused educational tools. The dataset contains approximately 5 million entries in arrow and Parquet formats, primarily in English.