edithgithinji/DeepScaleR-Preview-Dataset

Data Our training dataset consists of approximately 40,000 unique mathematics problem-answer pairs compiled from: AIME (American Invitational Mathematics Examination) problems (1984-2023) AMC (American Mathematics Competition) problems (prior to 2023) Omni-MATH dataset Still dataset Format Each row in the JSON dataset contains: problem: The mathematical question text, formatted with LaTeX notation. solution: Offical solution to the problem, including LaTeX… See the full description on the dataset page: https://huggingface.co/datasets/edithgithinji/DeepScaleR-Preview-Dataset.