Summary of Functional Homotopy: Smoothing Discrete Optimization Via Continuous Parameters For Llm Jailbreak Attacks, by Zi Wang et al.
Functional Homotopy: Smoothing Discrete Optimization via Continuous Parameters for LLM Jailbreak Attacks
by Zi Wang, Divyam Anshumaan, Ashish Hooda, Yudong Chen, Somesh Jha
First submitted to arxiv on: 5 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel optimization approach, called functional homotopy, to address the challenge of applying gradient-based techniques to language models. The method leverages the duality between model training and input generation by constructing a series of easy-to-hard optimization problems. The approach is applied to jailbreak attack synthesis for large language models (LLMs), achieving a significant improvement in success rate over existing methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper finds a way to make language models better at responding to tricky questions. Right now, there are techniques that help with image recognition, but they don’t work well with words. The researchers came up with a new approach called functional homotopy that helps the model generate more accurate and relevant responses. They tested it on large language models and found it worked much better than other methods. |
Keywords
* Artificial intelligence * Optimization
