- (Optional) Update conda:
conda update -n base -c defaults conda -y
- Create and activate the conda environment:
conda create -n origin_of_scaling_env python=3.11 -y && conda activate origin_of_scaling_env
- Install the required packages:
conda install -c conda-forge numpy pandas scipy matplotlib seaborn pyarrow fsspec datasets huggingface_hub -y
- If generating the Many-Shot In-Context Learning data:
conda install pytorch-gpu torchvision torchaudio pytorch-cuda=12.4 -c pytorch -c nvidia -y
pip install protobuf sentencepiece
To cite this work, please use:
@inproceedings{schaeffer2025monkeypowerlaws,
title={How Do Large Language Monkeys Get Their Power (Laws)?},
author={Rylan Schaeffer and Joshua Kazdan and John Hughes and Jordan Juravsky and Sara Price and Aengus Lynch and Erik Jones and Robert Kirk and Azalia Mirhoseini and Sanmi Koyejo},
booktitle={Forty-second International Conference on Machine Learning},
year={2025},
url={https://openreview.net/forum?id=QqVZ28qems}
}
Questions? Comments? Interested in collaborating? Open an issue or email rschaef@cs.stanford.edu and sanmi@cs.stanford.edu.