Dataset · amphora
Research-level mathematical problems in English
amphora/ResearchMath-14k
14,056 complex mathematical problems from research papers and open lists for model training and evaluation.
What's inside
The dataset contains 14,056 records of research-level mathematical problems in English. The materials are extracted from research papers, open problem lists, workshop proceedings, and other academic sources. Each record includes the original question, a rewritten self-contained problem statement, taxonomy labels, and the problem's open status.
What you can do with it
- Fine-tuning (SFT) language models for solving complex mathematical problems
- Testing AI agents' ability to reason about open scientific problems
- Classifying and structuring academic mathematical texts
- Tasks
- Text generation, Question answering
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 741
- Likes
- 57
How to load it
from datasets import load_dataset
ds = load_dataset("amphora/ResearchMath-14k")