README.md

metadata

language:
  - en
  - eu
license: mit
task_categories:
  - text-generation
  - question-answering
tags:
  - math
  - word-problems
  - maseu
  - mawps
  - asdiv_a
  - svamp
  - gsm8k
  - mgsm
  - multilingual
  - basque
  - low-resource
pretty_name: MASEU
dataset_info:
  features:
    - name: Question
      dtype: string
    - name: Numbers
      sequence: float64
    - name: Equation
      sequence: string
    - name: Answer
      dtype: float64
    - name: group_nums
      sequence: int64
    - name: Body
      dtype: string
    - name: Ques
      dtype: string
    - name: id
      dtype: int64
  splits:
    - name: train
      num_bytes: 118560
      num_examples: 195
    - name: test
      num_bytes: 816875
      num_examples: 1584
  download_size: 277198
  dataset_size: 935435
configs:
  - config_name: en
    data_files:
      - split: train
        path: en/train-*
      - split: test
        path: en/test-*
  - config_name: eu
    data_files:
      - split: train
        path: eu/train-*
      - split: test
        path: eu/test-*

MASEU Multilingual

Dataset Description

MASEU is a dataset specifically constructed to enable reliable and linguistically faithful evaluation of mathematical reasoning in Basque, a low-resource language. It is based on a manually curated subset of the mawps-asdiv-a_svamp corpus, which merges three well-established benchmarks in the domain of Math Word Problems (MWPs): MAWPS, ASDiv-A, and SVAMP. These datasets were selected for their diversity in reasoning types, consistent structure, and pedagogical value, making them particularly suitable for testing LLM performance in multilingual and instructional contexts.

The Basque portion of MASEU comprises 195 train entries and 1584 test entries, all carefully translated into Basque by a single native speaker, without the use of any machine translation tools or automated assistance. The translation process was fully manual and carefully controlled to ensure both mathematical fidelity and linguistic naturalness, faithfully preserving the original intent, difficulty level, and logical structure of each problem. This guarantees that the Basque version reflects idiomatic usage while maintaining conceptual equivalence, enabling robust reasoning evaluation without introducing semantic drift.

Available Configs

Config	Language
en	English
eu	Basque

Splits

Split	Examples	Description
train	195	Training set
test	1584	Test set for evaluation

Usage

from datasets import load_dataset

# Load the English config
dataset = load_dataset("inigomartinez/MASEU", name="en")

# Load the Basque config
dataset = load_dataset("inigomartinez/MASEU", name="eu")

# Access the splits
train = dataset["train"]
test = dataset["test"]

Data Fields

Field	Type	Description
id	int64	Unique identifier for the example
Question	string	Full math problem in natural language
Body	string	Body of the problem statement
Ques	string	Specific question of the problem
Numbers	sequence float64	List of relevant numbers present in the problem
Equation	sequence string	Equation(s) that solve the problem
Answer	float64	Final numerical answer
group_nums	sequence int64	Grouping of the numbers in the problem

Format

The data is stored in Parquet format.

Citation

@inproceedings{martinezcriado-etal-2026-benchmarking,
  title = {Benchmarking Mathematical Reasoning in a Low-Resource Language: Structured Prompting and Evaluation in Basque},
  author = {Martinez-Criado, Inigo and Soroa, Aitor and Barnes, Jeremy},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference (LREC 2026)},
  month = {May},
  year = {2026},
  pages = {5268--5289},
  address = {Palma, Mallorca, Spain},
  publisher = {European Language Resources Association (ELRA)},
  editor = {Piperidis, Stelios and Bel, Núria and van den Heuvel, Henk and Ide, Nancy and Krek, Simon and Toral, Antonio},
  doi = {10.63317/48gd2ysjuh8s},
  abstract = {Large Language Models (LLMs) have shown impressive performance on tasks requiring complex reasoning, but most evaluations tend to focus on English and other high-resource languages. This work investigates how well LLMs perform mathematical reasoning in low-resource languages, using Basque as a primary case study. To support this analysis, we introduce MASEU, a benchmark designed to evaluate reasoning in Basque across arithmetic, algebraic, and logical tasks. We then use this dataset to address three key questions: 1) how well do LLMs support Basque in reasoning tasks, 2) to what extent can including English in prompts improve results, and 3) what is the effect of continued pretraining in Basque? To explore these aspects, we use prompting strategies adapted for mathematical reasoning, building upon the foundations of CoT prompting and one of its subsequent evolutions, DUP prompting, which together allow for more precise experimentation across zero-shot and few-shot settings, providing insights into how multilingual models handle reasoning tasks in underrepresented languages.}
}

@inproceedings{koncel-kedziorski-etal-2016-mawps,
  title = {MAWPS: A Math Word Problem Repository},
  author = {Koncel-Kedziorski, Rik and Roy, Subhro and Amini, Aida and Kushman, Nate and Hajishirzi, Hannaneh},
  booktitle = {Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  year = {2016}
}

@inproceedings{miao-etal-2020-diverse,
  title = {A Diverse Corpus for Evaluating and Developing English Math Word Problem Solvers},
  author = {Miao, Shen-yun and Liang, Chao-Chun and Su, Keh-Yih},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2020},
  year = {2020}
}

@inproceedings{patel-etal-2021-nlp,
  title = {Are NLP Models really able to Solve Simple Math Word Problems?},
  author = {Patel, Arkil and Bhatt, Satwik and Baral, Chitta},
  booktitle = {Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  year = {2021}
}