LIMO (Less Is More for Reasoning) is a mathematical reasoning dataset that aims to train and evaluate the mathematical reasoning ability of large models by carefully selecting high-quality training samples.LIMO: Less is More for ReasoningThis dataset is mainly used to train the mathematical problem-solving ability of large models and improve their performance in mathematical exams and competition questions (such as AIME, MATH-500, etc.). The LIMO dataset is characterized by high quality, small scale, and high efficiency. The dataset contains only 817 high-quality mathematical reasoning samples, but it scored 44.5 in the AIME 2025 evaluation, which is close to the model trained with 800,000 samples.

Citation

@misc{ye2025limoreasoning,
title={LIMO: Less is More for Reasoning},
author={Yixin Ye and Zhen Huang and Yang Xiao and Ethan Chern and Shijie Xia and Pengfei Liu},
year={2025},
eprint={2502.03387},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2502.03387},
}