lucy3/aftermath_questions_to_categories

Aftermath of DrawEduMath

This contains questions_to_categories.json, for recreating the results of the paper titled "The Aftermath of DrawEduMath: Vision Language Models Underperform with Struggling Students and Misdiagnose Errors".

This file contains a mapping from DrawEduMath questions to question types, following the taxonomy introduced in the original DrawEduMath paper.

Please consult the datacard for DrawEduMath for detailed information about data source.

Quick links:

• Aftermath of DrawEduMath code: https://github.com/lucy3/aftermath_drawedumath/tree/main
• Aftermath of DrawEduMath paper: https://arxiv.org/abs/2603.00925

Example Use

import json
with open('questions_to_categories.json', 'r') as infile:
    question_to_category = json.load(infile)

License

This dataset is licensed under CC-BY-NC-4.0. It is intended for research and educational purposes following ASSISTments's Responsible Use Guidelines.

Citation

@misc{lucy2026aftermathdrawedumathvisionlanguage,
      title={The Aftermath of DrawEduMath: Vision Language Models Underperform with Struggling Students and Misdiagnose Errors}, 
      author={Li Lucy and Albert Zhang and Nathan Anderson and Ryan Knight and Kyle Lo},
      year={2026},
      eprint={2603.00925},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2603.00925}, 
}