Cumulative Reasoning with Large Language Models

Introduction

Official implementation of paper "Cumulative Reasoning with Large Language Models" (https://arxiv.org/abs/2308.04371).

• Achieving 98% accuracy for the Game of 24 (+24% compared to Tree-of-Thoughts)!

• Achieving 58% accuracy on the MATH dataset (+4.2% compared to PHP)!

• Achieving 43% relative improvement on the hardest Level 5 MATH problems (22.4% to 32.1%)!

Installation

pip install -r requirements.txt

For more usage help, please refer to the README.md in each subdirectory.

Acknowledgement

This repo is mainly based on Guidance, HuggingFace and Tree of Thoughts. Thanks for their wonderful work!

Citations

Please cite the paper and star this repo if you use Cumulative Reasoning (CR) and find it interesting/useful, thanks! Feel free to contact [email protected] | [email protected] or open an issue if you have any questions.

@article{zhang2023cumulative,
  title={Cumulative Reasoning With Large Language Models},
  author={Zhang, Yifan and Yang, Jingqin and Yuan, Yang and Yao, Andrew Chi-Chih},
  journal={arXiv preprint arXiv:2308.04371},
  year={2023}
}