README.md

---
language:
- en
license: mit
size_categories:
- 10K<n<100K
task_categories:
- text-generation
pretty_name: MATH
configs:
- config_name: numeric
  data_files:
  - split: train
    path: numeric/train-*
  - split: test
    path: numeric/test-*
- config_name: original
  data_files:
  - split: train
    path: original/train-*
  - split: test
    path: original/test-*
dataset_info:
- config_name: default
  features:
  - name: problem
    dtype: string
  - name: level
    dtype: string
  - name: type
    dtype: string
  - name: solution
    dtype: string
  - name: extracted_solution
    dtype: string
  splits:
  - name: train
    num_bytes: 6062403
    num_examples: 7500
  - name: test
    num_bytes: 3783919
    num_examples: 5000
  download_size: 4921628
  dataset_size: 9846322
- config_name: numeric
  features:
  - name: problem
    dtype: string
  - name: level
    dtype: string
  - name: type
    dtype: string
  - name: solution
    dtype: string
  - name: extracted_solution
    dtype: float64
  splits:
  - name: train
    num_bytes: 3712169
    num_examples: 4866
  - name: test
    num_bytes: 2229985
    num_examples: 3199
  download_size: 3035498
  dataset_size: 5942154
- config_name: original
  features:
  - name: problem
    dtype: string
  - name: level
    dtype: string
  - name: type
    dtype: string
  - name: solution
    dtype: string
  - name: extracted_solution
    dtype: string
  splits:
  - name: train
    num_bytes: 6062403
    num_examples: 7500
  - name: test
    num_bytes: 3783919
    num_examples: 5000
  download_size: 4921628
  dataset_size: 9846322
---
# Dataset Card for "competition_math"

Added column with final solution extracted from \boxed{} tags.
Added `numeric` congig that only contains questions with numeric answers.

## Dataset Description

- **Homepage:** https://github.com/hendrycks/math/blob/main/README.md
- **Repository:** https://github.com/hendrycks/math
- **Paper:** https://arxiv.org/abs/2103.03874

### Dataset Summary

MATH contains 12,500 challenging competition mathematics problems. Each problem in MATH has a full step-by-step solution which can be used to teach models to generate answer derivations and explanation
This dataset card aims to be a base template for new datasets.


### Languages

[English]

## Dataset Structure

### Data Instances

7 sub-datasets

### Data Splits

training: 7500
test: 5000


## Additional Information


### Licensing Information

MIT but check the [Legal Compliance](https://arxiv.org/pdf/2103.03874.pdf) section in appendix B of the paper as well as the [repo](https://github.com/hendrycks/math/blob/main/LICENSE).

### Citation Information

@article{hendrycksmath2021,
  title={Measuring Mathematical Problem Solving With the MATH Dataset},
  author={Dan Hendrycks and Collin Burns and Saurav Kadavath and Akul Arora and Steven Basart and Eric Tang and Dawn Song and Jacob Steinhardt},
  journal={NeurIPS},
  year={2021}
}