SPRI (ICML 2025)

This repo contains the dataset and code for our ICML 2025 paper. If you use the code or dataset, please cite our paper.

Title: SPRI: Aligning Large Language Models with Context-Situated Principles

Authors: Hongli Zhan, Muneeza Azmat, Raya Horesh, Junyi Jessy Li, Mikhail Yurochkin

@inproceedings{zhan2025spri,
  title={SPRI: Aligning Large Language Models with Context-Situated Principles},
  author={Zhan, Hongli and Azmat, Muneeza and Horesh, Raya and Li, Junyi Jessy and Yurochkin, Mikhail},
  booktitle={Proceedings of the 42nd International Conference on Machine Learning},
  year={2025}
}

Code for SPRI

$ conda create --name spri python==3.12.7
$ conda activate spri
$ pip install -r requirements.txt

SFT Data Generated Using SPRI

We additionally release the SFT datasets generated by SPRI on Huggingface, and you can find them here: https://huggingface.co/papers/2502.03397.

Repository Status

I'm actively putting up the documentation of the code. If you have any questions, feel free to email me at [email protected].