Skip to content

Latest commit

 

History

History
194 lines (141 loc) · 9.34 KB

README.md

File metadata and controls

194 lines (141 loc) · 9.34 KB

OpenUnlearning

An easily extensible framework unifying LLM unlearning evaluation benchmarks.


📖 Overview

We provide efficient and streamlined implementations of the TOFU, MUSE unlearning benchmarks while supporting 5 unlearning methods, 3+ datasets, 6+ evaluation metrics, and 7+ LLMs. Each of these can be easily extended to incorporate more variants.

We invite the LLM unlearning community to collaborate by adding new benchmarks, unlearning methods, datasets and evaluation metrics here to expand OpenUnlearning's features, gain feedback from wider usage and drive progress in the field.

⚠️ Notice (Updated: February 27, 2025)
This repository replaces the original TOFU codebase, which can be found at github.com/locuslab/tofu and isn't maintained anymore.

🗃️ Available Components

We provide several variants for each of the components in the unlearning pipeline.

Component Available Options
Benchmarks TOFU, MUSE
Unlearning Methods GradAscent, GradDiff, NPO, SimNPO, DPO
Evaluation Metrics Verbatim Probability, Verbatim ROUGE, QA-ROUGE, MIA Attacks, TruthRatio, Model Utility
Datasets MUSE-News (BBC), MUSE-Books (Harry Potter), TOFU (different splits)
Model Families LLaMA 3.2, LLaMA 3.1, LLaMA-2, Phi-3.5, ICLM (from MUSE), Phi-1.5, Gemma

📌 Table of Contents


⚡ Quickstart

🛠️ Environment Setup

conda create -n unlearning python=3.11
conda activate unlearning
pip install .
pip install --no-build-isolation flash-attn==2.6.3

💾 Data Setup

Download the log files containing metric results from the models used in the supported benchmarks (including the retain model logs used to compare the unlearned models against).

python setup_data.py # populates saves/eval with evaluation results of the uploaded models

🧪 Running Experiments

We provide an easily configurable interface for running evaluations by leveraging Hydra configs. For a more detailed documentation of aspects like running experiments, commonly overriden arguments, interfacing with configurations, distributed training and simple finetuning of models, refer docs/experiments.md.

🚀 Perform Unlearning

An example command for launching an unlearning process with GradAscent on the TOFU forget10 split:

python src/train.py --config-name=unlearn.yaml experiment=unlearn/tofu/default \
  forget_split=forget10 retain_split=retain90 trainer=GradAscent

📊 Perform an Evaluation

An example command for launching a TOFU evaluation process on forget10 split:

python src/eval.py --config-name=eval.yaml experiment=eval/tofu/default \
  model=Llama-3.2-1B-Instruct \
  model.model_args.pretrained_model_name_or_path=open-unlearning/tofu_Llama-3.2-1B-Instruct_full
  • experiment-Path to the evaluation configuration configs/experiment/eval/tofu/default.yaml.
  • model- Sets up the model and tokenizer configs for the Llama-3.2-1B-Instruct model.
  • model.model_args.pretrained_model_name_or_path- Overrides the default experiment config to evaluate a model from a HuggingFace ID (can use a local model checkpoint path as well).

For more details about creating and running evaluations, refer docs/evaluation.md.

📜 Running Baseline Experiments

The scripts below execute standard baseline unlearning experiments on the TOFU and MUSE datasets, evaluated using their corresponding benchmarks. The expected results for these are in docs/results.md.

bash scripts/tofu_unlearn.sh
bash scripts/muse_unlearn.sh

➕ How to Add New Components

Adding a new component (trainer, evaluation metric, benchmark, model, or dataset) requires defining a new class, registering it, and creating a configuration file. Learn more about adding new components in docs/components.md.

Please feel free to raise a pull request for any new features after setting up the environment in development mode.

pip install .[flash-attn, dev]

📚 Further Documentation

For more in-depth information on specific aspects of the framework, refer to the following documents:

Documentation Contains
docs/components.md Instructions on how to add new components such as trainers, benchmarks, metrics, models, datasets, etc.
docs/evaluation.md Detailed instructions on creating and running evaluation metrics and benchmarks.
docs/experiments.md Guide on running experiments in various configurations and settings, including distributed training, fine-tuning, and overriding arguments.
docs/hydra.md Explanation of the Hydra features used in configuration management for experiments.
docs/results.md Reference results from various unlearning methods run using this framework on TOFU and MUSE benchmarks.

🔗 Support & Contributors

Developed and maintained by Vineeth Dorna (@Dornavineeth) and Anmol Mekala (@molereddy).

If you encounter any issues or have questions, feel free to raise an issue in the repository 🛠️.

📝 Citation

This repo is inspired from LLaMA-Factory. We acknowledge the TOFU and MUSE benchmarks, which served as the foundation for our re-implementation.


If you use OpenUnlearning in your research, please cite:

@misc{openunlearning2025,
  title={OpenUnlearning: A Unified Framework for LLM Unlearning Benchmarks},
  author={Dorna, Vineeth and Mekala, Anmol and Zhao, Wenlong and McCallum, Andrew and Kolter, J Zico and Maini, Pratyush},
  year={2025},
  howpublished={\url{https://github.com/locuslab/open-unlearning}},
  note={Accessed: February 27, 2025}
}
@inproceedings{maini2024tofu,
  title={TOFU: A Task of Fictitious Unlearning for LLMs},
  author={Maini, Pratyush and Feng, Zhili and Schwarzschild, Avi and Lipton, Zachary Chase and Kolter, J Zico},
  booktitle={First Conference on Language Modeling},
  year={2024}
}
To cite other benchmarks used from OpenUnlearning
@article{shi2024muse,
  title={Muse: Machine unlearning six-way evaluation for language models},
  author={Shi, Weijia and Lee, Jaechan and Huang, Yangsibo and Malladi, Sadhika and Zhao, Jieyu and Holtzman, Ari and Liu, Daogao and Zettlemoyer, Luke and Smith, Noah A and Zhang, Chiyuan},
  journal={arXiv preprint arXiv:2407.06460},
  year={2024}
}

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.