Wenjie Qu, National University of Singapore; Yijun Sun, Hong Kong University of Science and Technology; Xuanming Liu, Tao Lu, and Yanpei Guo, National University of Singapore; Kai Chen, Hong Kong University of Science and Technology; Jiaheng Zhang, National University of Singapore
Large Language Models (LLMs) are widely employed for their ability to generate human-like text. However, service providers may deploy smaller models to reduce costs, potentially deceiving users. Zero-Knowledge Proofs (ZKPs) offer a solution by allowing providers to prove LLM inference without compromising the privacy of model parameters. Existing solutions either do not support LLM architectures or suffer from significant inefficiency and tremendous overhead. To address this issue, this paper introduces several new techniques. We propose new methods to efficiently prove linear and non-linear layers in LLMs, reducing computation overhead by orders of magnitude. To further enhance efficiency, we propose constraint fusion to reduce the overhead of proving non-linear layers and circuit squeeze to improve parallelism. We implement our efficient protocol, specifically tailored for popular LLM architectures like GPT-2, and deploy optimizations to enhance performance. Experiments show that our scheme can prove GPT-2 inference in less than 25 seconds. Compared with state-of-the-art systems such as Hao et al. (USENIX Security'24) and ZKML (Eurosys'24), our work achieves nearly 279x and 185x speedup, respectively.
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.
