metadata
dataset_info:
features:
- name: id
dtype: int64
- name: problem
dtype: string
- name: solution
dtype: string
- name: answer
dtype: string
- name: problem_type
dtype: string
- name: question_type
dtype: string
- name: problem_is_valid
dtype: string
- name: solution_is_valid
dtype: string
- name: source
dtype: string
- name: synthetic
dtype: bool
splits:
- name: train
num_bytes: 128094696
num_examples: 75580
download_size: 57263310
dataset_size: 128094696
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
This dataset is a curated subset of the original AI-MO/NuminaMath-1.5 dataset. It has been filtered to create a high-quality corpus of human-authored, validated math problems suitable for training and evaluating language models on mathematical reasoning.
Filtering Criteria
This subset was created by applying the following conditions to the 'train' split of the original dataset:
- Valid Problems: Includes only problems marked as complete and well-formed.
- Human-Authored Content Only: Excludes all synthetically generated problems.
- Cleaned Text: Removed all entries containing URLs, file links, or markdown images.
Data Fields
The data fields are inherited from the original dataset and include:
problem: The mathematical problem statement in LaTeX.solution: A step-by-step, Chain-of-Thought style solution.answer: The final answer to the problem.problem_type: The mathematical domain (e.g., Algebra, Geometry).question_type: The style of the problem (e.g., proof, math-word-problem).source: The origin of the problem (e.g., olympiads, cn_k12).