Datasetpublic

Corpus-NZ/Math-Mini

Clean Math Reasoning Dataset A clean dataset for training and evaluating language models on mathematical problem solving. The dataset contains concise mathematical question-and-answer pairs designed to improve model performance on structured numerical reasoning tasks. Dataset Structure The dataset contains two fields: Column Description prompt A mathematical problem or question response The corresponding solution Example: prompt: 48392+92831=?… See the full description on the dataset page: https://huggingface.co/datasets/Corpus-NZ/Math-Mini.

sourceHugging Faceotherupdated 28d agoView on Hugging Face

0likes38downloads

Dataset Card

Clean Math Reasoning Dataset

A clean dataset for training and evaluating language models on mathematical problem solving.

The dataset contains concise mathematical question-and-answer pairs designed to improve model performance on structured numerical reasoning tasks.

Dataset Structure

The dataset contains two fields:

Column	Description
`prompt`	A mathematical problem or question
`response`	The corresponding solution

Example:

text

prompt:
48392+92831=?

response:
48392+92831=141223

Tasks Covered

The dataset includes examples across several mathematical areas:

—Arithmetic operations
—Algebra
—Fractions
—Percentages
—Powers and exponents
—Equations
—Geometry calculations

Features

—Clean CSV format
—Consistent input-output structure
—No missing values
—Compact and efficient formatting
—Suitable for language model training and evaluation

Intended Uses

This dataset can be used for:

—Supervised fine-tuning
—Mathematical reasoning experiments
—Model evaluation
—Educational AI systems
—Benchmarking arithmetic capabilities

Data Format

CSV format:

csv

prompt,response
"12^5=?","12^5=248832"
"25% of 8400=?","8400*25/100=2100"
"Solve 7x+9=72","7x=63, x=9"

Limitations

This dataset focuses on structured mathematical tasks. It is intended to improve performance on numerical reasoning patterns and does not cover advanced mathematical topics such as formal proofs, abstract mathematics, or complex real-world modeling.

License

Open Data Attribution Training Disclosure License (ODATL‑1.0)