Optimization methods are widely employed in deep learning to identify and mitigate undesired model responses. While gradient-based techniques have proven effective for image models, their application to language models is hindered by the discrete nature of the input space. This study introduces a novel optimization approach, termed the \emph{functional homotopy} method, which leverages the functional duality between model training and input generation. By constructing a series of easy-to-hard optimization problems, we iteratively solve these problems using principles derived from established homotopy methods. We apply this approach to jailbreak attack synthesis for large language models (LLMs), achieving a improvement in success rate over existing methods in circumventing established safe open-source models such as Llama-2 and Llama-3.
View on arXiv@article{wang2025_2410.04234,
title={ Functional Homotopy: Smoothing Discrete Optimization via Continuous Parameters for LLM Jailbreak Attacks },
author={ Zi Wang and Divyam Anshumaan and Ashish Hooda and Yudong Chen and Somesh Jha },
journal={arXiv preprint arXiv:2410.04234},
year={ 2025 }
}