Skip to content

apartresearch/Verified_addition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 

Repository files navigation

Official Repository for Increasing Trust in Language Models through the Reuse of Verified Circuits

Authors Philip Quirke, Clement Neo and Fazl Barez

Understanding the inner workings of machine learning models like Transformers is vital for their safe and ethical use. This repository contains Colabs that train addition models, subtraction models, addition and subtraction models. The resulting models have very low loss and can correctly predict one million succcessive questions.

The associated paper details the algorithm of the addition algorithm:

The Accurate_Math_Train.ipynb supports the optional re-use of one model when training a new model:

Our initial (less capable) Colabs:

Our final (more-advanced) Colabs:

  • train a n-digit integer addition model, or subtraction model, or "addition and subtraction" model ( Accurate_Math_Train.ipynb )
  • allow the untrained "addition and subtraction" model to be initialised with a trained "addition" model as an experiment in model re-use
  • analyse the trained models using to provide insights into their algorithms ( Accurate_Math_Analyse.ipynb )

The Accurate_Math_Train.ipynb Colab is preconfigured to train 9 different models (with 5 or 6 digits, 2 or 3 layers, 3 or 4 heads, doing addition, subtraction or both). A one-line code change in "Part 1B" swaps between these models. Training of one model can take ~30 minutes. The trained model weights (*.pth) are saved to a public HuggingFace folder.

The Accurate_Math_Analyse.ipynb Colab is preconfigured to analyse 7 different models (with 5 or 6 digits, 2 or 3 layers, 3 or 4 heads, doing addition, subtraction or both). A one-line code change in "Part 1B" swaps between these models. For speed, this CoLab loads the selected model weights from the public HuggingFace folder, allowing the analysis to start promptly. The Colab produces various graphs displaying its analysis.

Reference

If you use our work, please consider citing:

@misc{quirke2024increasing,
      title={Increasing Trust in Language Models through the Reuse of Verified Circuits}, 
      author={Philip Quirke and Clement Neo and Fazl Barez},
      year={2024},
      eprint={2402.02619},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published