NSmark: Null Space Based Black-box Watermarking Defense Framework for Pre-trained Language Models

Official Code for Paper "NSmark: Null Space Based Black-box Watermarking Defense Framework for Pre-trained Language Models".link

Environment Setup

To set up the environment, please refer to requirements.txt in python environment.

Configuration

The configs folder contains configuration files in YAML format. You will need to modify these files according to your needs.

Data Preparation

Place the datasets mentioned in the YAML configuration files into the data folder. For specific paths where the datasets should be stored, refer to the __init__.py file within the data folder.

Model Storage

Refer to the YAML configuration files and store the models involved, such as BERT, in a folder of your choice.

Main Script

The nsmark_watermarking.py file serves as the main entry point for the program.

Running the Script

Modify the args in run.sh to run this script.

Trigger Selection

To bind the trigger words with user information, you can refer to sign.py. For convenience, trigger word is specified for experimentation.

Citing NSmark

@article{zhao2024nsmarknullspacebased,
  title={NSmark: Null Space Based Black-box Watermarking Defense Framework for Pre-trained Language Models},
  author={Zhao, Haodong and Hu, Jinming and Li, Peixuan and Li, Fangqi and Sha, Jinrui and Chen, Peixuan and Zhang, Zhuosheng and Liu, Gongshen},
  journal={arXiv preprint arXiv:2410.13907},
  year={2024}
}

License

This project is licensed under the Apache-2.0 License.