Anon13576245/paper-data

Paper Data

This dataset is a merged collection of mathematical problem data used for evaluation. It combines final-answer problems, proof-writing problems, refutation-style problems, image-only multiple-choice problems, and Lean formalization problems into one public dataset repository.

The rows come from multiple competition or benchmark splits. Because the source formats differ, many columns are intentionally nullable. A null value means that the field does not apply to that row or was not available in the source data.

Included Splits

The competition column identifies the source split for each row. The combined upload includes:

• kangaroo/kangaroo_2025_1-2
• kangaroo/kangaroo_2025_3-4
• kangaroo/kangaroo_2025_5-6
• kangaroo/kangaroo_2025_7-8
• kangaroo/kangaroo_2025_9-10
• kangaroo/kangaroo_2025_11-12
• aime/aime_2026
• hmmt/hmmt_feb_2026
• arxivlean/march
• arxiv/december
• arxiv/january
• arxiv/february
• arxiv/march
• arxiv_false/february
• arxiv_false/march
• usamo/usamo_2026
• putnam/putnam_2025
• imc/imc_2025
• miklos/2025
• apex/apex_2025
• apex/shortlist_2025

Schema

Column	Type	Description
competition	string	Source split identifier. Use this together with problem_idx as the stable row key.
problem_idx	integer	Original problem index within the source split.
problem	string, nullable	Problem statement for text-based rows, usually in LaTeX or plain text. Null for image-only rows.
image	image, nullable	Problem image for image-only rows, currently used by the Kangaroo splits. Null for text-based rows.
file_name	string, nullable	File name for rows with an image asset.
answer	string, nullable	Final answer for final-answer tasks. For Lean rows, this contains the formal statement. Null for proof-writing or rubric-graded rows where no final-answer field is used.
formal_statement	string, nullable	Lean formal statement for Lean formalization rows.
problem_type	list of strings, nullable	Topic or category labels when available.
source	string, nullable	Source note or source identifier when available.
title	string, nullable	Source title metadata when available, primarily for arXiv-derived rows.
authors	string, nullable	Source author metadata when available, primarily for arXiv-derived rows. This refers to source material authors, not dataset authors.
points	float, nullable	Maximum score for rubric-graded proof-writing rows.
grading_scheme	string, nullable	Rubric, grading criteria, or judge-facing scheme for proof-writing and refutation-style rows. Some nested source values are serialized as JSON strings for a stable dataset schema.
sample_solution	string, nullable	Reference or sample solution where available.

Row Types

Final-answer rows contain a text problem and an answer. These include sources such as AIME, HMMT, ArXivMath, Apex, and Kangaroo answer labels. For Kangaroo rows, the problem statement is stored in the image column rather than problem.

Proof-writing rows contain a problem, a grading_scheme, and usually points. They may also include sample_solution. These rows generally have answer = null because they are evaluated by rubric rather than exact answer matching.

Lean rows contain a natural-language problem, a Lean formal_statement, and the same formal statement in answer for compatibility with final-answer style loaders.

Refutation-style rows contain statements that are intended for proof or disproof evaluation. The grading_scheme describes how responses should be judged.

Loading Example

from datasets import load_dataset

dataset = load_dataset("Anon13576245/paper-data", split="train")

row = dataset[0]
print(row["competition"], row["problem_idx"])
print(row["problem"] or row["file_name"])

Notes

• The dataset is intentionally heterogeneous. Consumers should filter by competition or by non-null columns before applying task-specific evaluation logic.
• Null values are expected and should not be treated as missing data errors.
• Image rows can be accessed through the Hugging Face Image feature in the image column.
• The dataset card is identity-neutral. Metadata fields such as authors describe original source material where provided.