Neural Interactive Proofs Documentation

arXiv | OpenReview | GitHub | Website

This is the documentation of the experiments for the ICLR 2025 paper “Neural Interactive Proofs” [HAD25]. The source code is available on GitHub at SamAdamDay/neural-interactive-proofs.

The codebase is designed to be easy to use and extend. The guides section contains tutorials on how to use the codebase, and the reference section contains a comprehensive API reference.

Quickstart

To get experiments running fast, follow the steps below:

1. Install the package by following the installation instructions. Alternatively, build and run a Docker container by following the Docker instructions.

2. Create an experiment script by following the guide to running experiments. A basic example is:

   from nip import HyperParameters, run_experiment
   
   hyper_params = HyperParameters(
      scenario="code_validation",
      trainer="pure_text_ei",
      dataset="lrhammond/buggy-apps",
      interaction_protocol="nip",
   )
   
   run_experiment(hyper_params)
   

Contents

• Guides
  • Installation
  • Running Experiments
  • Contributing
  • Creating a New Protocol
  • Using a Docker Image
• API Reference
  • Public API
  • Developer API
• Changelog
  • [1.0.0] - 2025-03-10
• Glossary
• Bibliography

Citation

@inproceedings{neural_interactive_proofs,
   author = {Lewis Hammond and Sam Adam-Day},
   title = {Neural Interactive Proofs},
   booktitle = {The Thirteenth International Conference on Learning Representations (ICLR)},
   year = {2025},
   eprint={2412.08897},
   archivePrefix={arXiv},
   primaryClass={cs.AI},
}