Demonstration only. Corollary is a non-commercial research demo — no warranty, no support, not for commercial use or clinical decisions. Read the disclaimer

All datasets

Competition Math · AI-MO

olympiads-ref-base

A reference dataset of mathematical olympiad problems used as a foundation for NuminaMath training splits.

What olympiads-ref-base is

olympiads-ref-base is a canonical reference dataset compiled by AI-MO. It brings together problem sets from international and national mathematical olympiads within the domain of competition mathematics.

The dataset serves as the base collection from which downstream training splits for NuminaMath are generated, providing researchers with standard text records for mathematical machine learning tasks.

How it was made

The dataset is curated from international and national mathematical olympiad competitions and formatted as text entries.

Key facts

Publisher
AI-MO
Field
Mathematics
Category
Competition Math
Data format
Parquet
Modality
Text
Creation date
20 December 2024

What you can use it for

Deriving NuminaMath training splits
Provides the foundational problem corpus used to assemble downstream training subsets for NuminaMath.
Competition mathematics research
Supplies a standardised collection of national and international olympiad problems for mathematical reasoning studies.
Text-based mathematical problem analysis
Facilitates computational examination and data processing of mathematical problem formulations across tabular libraries such as Pandas and Polars.

Explore the data

No licence declared

The authors have not published a licence, which means no permission to redistribute the data has been granted to anyone. The rows are not served here; read them on the original dataset page.

How the columns are distributed

Over 13,087 rows of default/train

answer

string_text
min
1
median
5
mean
9.70
max
250

0.29% empty

problem

string_text
min
14
median
289
mean
332.90
max
10,355

problem_is_valid

string_label
  • Yes98%
  • Incomplete2%
  • More than one problem0%

problem_raw

string_text
min
14
median
289
mean
334.41
max
10,355

problem_type

string_label
  • Geometry30%
  • Combinatorics24%
  • Number Theory21%

question_type

string_label
  • math-word-problem58%
  • proof42%
  • MCQ0%

solution

string_text
min
1
median
983
mean
1325.82
max
24,064

solution_is_valid

string_label
  • Yes84%
  • Incomplete10%
  • Problem not solved6%

solution_raw

string_text
min
1
median
986
mean
1334.86
max
24,064

Strengths and limits

Good at

  • Acts as a canonical reference collection of international and national olympiad problems.
  • Delivered in efficient Parquet format for compatibility with modern data processing libraries such as Polars and Datasets.

Watch out for

  • Restricted to text modality and does not incorporate other data types.

Questions

Can I look inside olympiads-ref-base without downloading it?

Not here: it declares no licence, so no permission to redistribute its rows has been granted. Its description and statistics are still shown.

What licence is olympiads-ref-base under?

Its authors have not declared a licence, which means no permission to reuse or redistribute it has been granted.

What content does olympiads-ref-base contain?

It contains text records of international and national mathematical olympiad competition problems.

How is this dataset related to NuminaMath?

It acts as the canonical reference base from which downstream NuminaMath training splits are derived.

Written from olympiads-ref-base's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.