aieng-lab / MathBERT-mamut

huggingface.co
Total runs: 123
24-hour runs: 29
7-day runs: 30
30-day runs: 50
Model's Last Updated: July 08 2025

Introduction of MathBERT-mamut

Model Details of MathBERT-mamut

MAMUT-MathBert (Math Mutator MathBERT)

MAMUT-MathBERT is a pretrained language model based on tbs17/MathBERT , further pretrained on mathematical texts and formulas. It was introduced in MAMUT: A Novel Framework for Modifying Mathematical Formulas for the Generation of Specialized Datasets for Language Model Training .

Despite its base model is already a mathematical model, our training aims to improve the mathematical understanding even further, as shown in our paper.

Model Details
Overview

MAMUT-MPBERT was pretrained on four math-specific tasks across four datasets.

  • Mathematical Formulas (MF) : A Masked Language Modeling (MLM) task on math formulas written in LaTeX.
  • Mathematical Texts (MT) : An MLM task on natural language text containing inline LaTeX math ( mathematical texts ). The masking probability was biased toward mathematical tokens (inside math environment $...$) and domain-specific terms (e.g., sum , one , ...)
  • Named Math Formulas (NMF) : A Next-Sentence-Prediction (NSP)-style task: given a formula and the name of a mathematical identity (e.g., Pythagorean Theorem), classify whether they match.
  • Math Formula Retrieval (MFR) : Another NSP-style task to decide if two formulas describe the same mathematical identity or concept.

Training Overview

Model Sources
Uses

MAMUT-MathBERT is intended for downstream tasks that require improved mathematical understanding, such as:

  • Formula classification
  • Retrieval of semantically similar formulas
  • Math-related question answering

Note: This model was saved without the MLM or NSP heads and requires fine-tuning before use in downstream tasks.

Similarly trained models are MAMUT-BERT based on bert-base-cased and MAMUT-MPBERT based on AnReu/math_structure_bert (best of the three models according to our evaluation).

Training Details

Training configurations are described in Appendix C of the MAMUT paper .

Evaluation

The model is evaluated in Section 7 and Appendix C.4 of the MAMUT paper (MAMUT-MPBERT).

Environmental Impact
  • Hardware Type: 8xA100
  • Hours used: 48
  • Compute Region: Germany
Citation

BibTeX:

@misc{drechsel2025mamutnovelframeworkmodifying,
      title={{MAMUT}: A Novel Framework for Modifying Mathematical Formulas for the Generation of Specialized Datasets for Language Model Training}, 
      author={Jonathan Drechsel and Anja Reusch and Steffen Herbold},
      year={2025},
      eprint={2502.20855},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.20855}, 
}

Runs of aieng-lab MathBERT-mamut on huggingface.co

123
Total runs
29
24-hour runs
31
3-day runs
30
7-day runs
50
30-day runs

More Information About MathBERT-mamut huggingface.co Model

More MathBERT-mamut license Visit here:

https://choosealicense.com/licenses/apache-2.0

MathBERT-mamut huggingface.co

MathBERT-mamut huggingface.co is an AI model on huggingface.co that provides MathBERT-mamut's model effect (), which can be used instantly with this aieng-lab MathBERT-mamut model. huggingface.co supports a free trial of the MathBERT-mamut model, and also provides paid use of the MathBERT-mamut. Support call MathBERT-mamut model through api, including Node.js, Python, http.

MathBERT-mamut huggingface.co Url

https://huggingface.co/aieng-lab/MathBERT-mamut

aieng-lab MathBERT-mamut online free

MathBERT-mamut huggingface.co is an online trial and call api platform, which integrates MathBERT-mamut's modeling effects, including api services, and provides a free online trial of MathBERT-mamut, you can try MathBERT-mamut online for free by clicking the link below.

aieng-lab MathBERT-mamut online free url in huggingface.co:

https://huggingface.co/aieng-lab/MathBERT-mamut

MathBERT-mamut install

MathBERT-mamut is an open source model from GitHub that offers a free installation service, and any user can find MathBERT-mamut on GitHub to install. At the same time, huggingface.co provides the effect of MathBERT-mamut install, users can directly use MathBERT-mamut installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

MathBERT-mamut install url in huggingface.co:

https://huggingface.co/aieng-lab/MathBERT-mamut

Url of MathBERT-mamut

MathBERT-mamut huggingface.co Url

Provider of MathBERT-mamut huggingface.co

aieng-lab
ORGANIZATIONS

Other API from aieng-lab