metadata

license: mit
task_categories:
  - question-answering
language:
  - en
tags:
  - olympiad
  - mathematics
  - competition
size_categories:
  - 1K<n<10K

OlympiadBench (Split Version)

This dataset is a split version of the original knoveleng/OlympiadBench dataset.

Dataset Description

This dataset contains mathematical olympiad problems with their solutions, split into training and test sets.

Dataset Structure

Train split: 575 examples
Test split: 100 examples (last 100 examples from original dataset)

Data Fields

question: The olympiad problem statement
answer: The solution to the problem

Usage

from datasets import load_dataset

# Load the entire dataset
dataset = load_dataset("weijiezz/OlympiadBench-split")

# Load specific splits
train_dataset = load_dataset("weijiezz/OlympiadBench-split", split="train")
test_dataset = load_dataset("weijiezz/OlympiadBench-split", split="test")

Source

This dataset is derived from knoveleng/OlympiadBench.

License

MIT License (following the original dataset)