Exposing Jailbreak Vulnerabilities in LLM Applications with ARTKIT
Automated prompt-based testing to extract passwords from the Gandalf Challenge's LLM system
Link to article: https://towardsdatascience.com/exposing-jailbreak-vulnerabilities-in-llm-applications-with-artkit-d2df5f56ece8
Background
- As large language models (LLMs) become more widely adopted across different industries and domains, significant security risks have emerged and intensified. Several of these key concerns include breaches of data privacy, the potential for biases, and the risk of information manipulation.
- Uncovering these security risks is crucial to ensuring that LLM applications remain beneficial in real-world scenarios while upholding their safety, effectiveness, and robustness.
- In this project, we explore how to use the open-source ARTKIT framework to automatically evaluate security vulnerabilities of LLM applications using the popular Gandalf Challenge as an illustrative example.
Files
gandalf_challenge.ipynb: Jupyter notebook containing the codes for the walkthrough
References
- Official ARTKIT GitHub Repo
- Play the Gandalf Challenge
Acknowledgements
- Special thanks to Sean Anggani, Andy Moon, Matthew Wong, Randi Griffin, and Andrea Gao!
