---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: subfield
    dtype: string
  - name: context
    dtype: 'null'
  - name: problem
    dtype: string
  - name: solution
    dtype: string
  - name: final_answer
    sequence: string
  - name: is_multiple_answer
    dtype: bool
  - name: unit
    dtype: string
  - name: answer_type
    dtype: string
  - name: error
    dtype: string
  - name: original_solution
    sequence: string
  splits:
  - name: all_data
    num_bytes: 3030627
    num_examples: 825
  - name: train
    num_bytes: 2571441.090909091
    num_examples: 700
  - name: test
    num_bytes: 459185.9090909091
    num_examples: 125
  download_size: 2837889
  dataset_size: 6061254.0
configs:
- config_name: default
  data_files:
  - split: all_data
    path: data/all_data-*
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
extra_gated_prompt: 'By requesting access to this dataset, you agree to cite the following
  works in any publications or projects that utilize this data:

  - Putnam-AXIOM dataset: @article{putnam_axiom2024, title={Putnam-AXIOM: A Functional
  and Static Benchmark for Measuring Higher Level Mathematical Reasoning}, author={Aryan
  Gulati and Brando Miranda and Eric Chen and Emily Xia and Kai Fronsdal and Bruno
  de Moraes Dumont and Sanmi Koyejo}, journal={38th Conference on Neural Information
  Processing Systems (NeurIPS 2024) Workshop on MATH-AI}, year={2024}, url={https://openreview.net/pdf?id=YXnwlZe0yf},
  note={Preprint available at: https://openreview.net/pdf?id=YXnwlZe0yf}}

  - OlympiadBench dataset (if applicable): @misc{he2024olympiadbench, title={OlympiadBench:
  A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal
  Scientific Problems}, author={Chaoqun He and Renjie Luo and others}, year={2024},
  eprint={2402.14008}, archivePrefix={arXiv}, primaryClass={cs.CL}}'
---

# OlympiadBench Data set used in the Putnam-AXIOM Paper

The *Putnam-AXIOM* dataset is a benchmark for measuring advanced mathematical reasoning in large language models (LLMs). 
It includes challenging mathematical problems from the William Lowell Putnam Mathematical Competition, with both original problems and functional variations to address data contamination. The dataset aims to provide rigorous evaluations by requiring models to answer in boxed format, simplifying automatic answer matching.
OlympiadBench is an Olympiad level Benchmark with IMO questions.
We added the latex box (`\\boxed{$\\frac{1}{2}$}"`) for compatibility with Hendrycks's standard evaluations.

## Dataset Details

- **Title**: OlympiadBench: A Challenging Benchmark for Promoting AGI with
Olympiad-Level Bilingual Multimodal Scientific Problems
- **Paper**: Available at [arxiv](https://arxiv.org/avs/2402.14008)
- **Number of Problems**: 825

TODO: transform image into Asymptote language like this https://arxiv.org/pdf/2103.03874

## Usage

The dataset can be loaded in Python with the following code:

```python
from datasets import load_dataset
dataset = load_dataset("brando/olympiad-bench-imo-math-boxed-21-08-02024-v2", split='all')
```

Legal Compliance
TODO

Citation
If you use this dataset, please cite:

```bibtex
Copy code
@article{putnam_axiom2024,
  title = {Putnam-AXIOM: A Functional and Static Benchmark for Measuring Higher Level Mathematical Reasoning},
  author = {Aryan Gulati and Brando Miranda and Eric Chen and Emily Xia and Kai Fronsdal and Bruno de Moraes Dumont and Sanmi Koyejo},
  journal = {38th Conference on Neural Information Processing Systems (NeurIPS 2024) Workshop on MATH-AI},
  year = {2024},
  url = {https://openreview.net/pdf?id=YXnwlZe0yf},
  note = {Preprint available at: https://openreview.net/pdf?id=YXnwlZe0yf}
}
```
Cite the original work please:
```bibtex
@misc{he2024olympiadbench,
      title={OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems}, 
      author={Chaoqun He and Renjie Luo and Yuzhuo Bai and Shengding Hu and Zhen Leng Thai and Junhao Shen and Jinyi Hu and Xu Han and Yujie Huang and Yuxiang Zhang and Jie Liu and Lei Qi and Zhiyuan Liu and Maosong Sun},
      year={2024},
      eprint={2402.14008},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```