PrivacyRestore: Privacy-Preserving Inference in Large Language Models via Privacy Removal and Restoration
This addresses privacy concerns for users of online LLM services, but it is incremental as it builds on existing privacy protection methods.
The paper tackles the problem of protecting private information in user inputs during online LLM inference by proposing PrivacyRestore, a plug-and-play method that uses privacy removal and restoration via activation steering, and it shows effective privacy protection with acceptable performance and inference overhead on medical and legal datasets.
The widespread usage of online Large Language Models (LLMs) inference services has raised significant privacy concerns about the potential exposure of private information in user inputs to malicious eavesdroppers. Existing privacy protection methods for LLMs suffer from either insufficient privacy protection, performance degradation, or large inference time overhead. To address these limitations, we propose PrivacyRestore, a plug-and-play method to protect the privacy of user inputs during LLM inference. The server first trains restoration vectors for each privacy span and then release to clients. Privacy span is defined as a contiguous sequence of tokens within a text that contain private information. The client then aggregate restoration vectors of all privacy spans in the input into a single meta restoration vector which is later sent to the server side along with the input without privacy spans.The private information is restored via activation steering during inference. Furthermore, we prove that PrivacyRestore inherently prevents the linear growth of the privacy budget.We create three datasets, covering medical and legal domains, to evaluate the effectiveness of privacy preserving methods. The experimental results show that PrivacyRestore effectively protects private information and maintain acceptable levels of performance and inference overhead.