diff --git a/README.md b/README.md index 89510fa..8f31588 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ -# PIG: Privacy Jailbreak Attack on LLMs via Gradient-based Iterative In-Context Optimization +# PIG: Privacy Jailbreak Attack on LLMs via Gradient-based Iterative In-Context Optimization -> This repository contains the official code implementation of our paper: [![arXiv: paper](https://img.shields.io/badge/arXiv-paper-red.svg)](https://arxiv.org/abs/xxx) +> This repository contains the official code implementation of our paper: [![arXiv: paper](https://img.shields.io/badge/arXiv-paper-red.svg)](https://arxiv.org/abs/2505.6437443) ![PIG](./img/PIG.png) @@ -21,16 +21,16 @@ pip install -r requirements.txt ## Datasets -You can download the Enron Email dataset and TrustLLM dataset [here](https://drive.google.com/drive/folders/16Th72F_QcxRAryOIk9L2t0oIps1xnHGW) and place them under the `./data` directory. +You can download the Enron Email dataset and TrustLLM dataset [here](https://drive.google.com/drive/folders/16Th72F_QcxRAryOIk9L2t0oIps1xnHGW) and place them under the `./data`. ## Usage You can run a privacy jailbreak attack using the following steps: -1. First, modify parameters such as `dataset`, `target_model_name`, `attack_model_name`, or `eval_model_name` in script `run.sh`. -2. Then, execute the privacy jailbreak attack by running `bash run.sh`. Use the `tail` command to monitor the `log` file in real time. +1. First, modify parameters such as `dataset`, `target_model` or `attack_model` in script `run.sh`. +2. Then, execute the privacy jailbreak attack by running `bash run.sh`. 3. Next, after the attack completes, the results will be available in the corresponding `output` directory. -4. Finally, evaluate the results using `python eval.py` to compute various metrics such as the Attack Success Rate (ASR). +4. Finally, evaluate the results using `python eval.py` to compute various metrics such as the ASR. ## Acknowledgements