Proactive Privacy Amnesia for Large Language Models: Safeguarding PII with Negligible Impact on Model Utility
This addresses privacy risks for users of large language models, offering a novel method to balance protection and utility, though it is incremental in improving existing defenses.
The paper tackles the problem of protecting personally identifiable information (PII) in large language models from malicious attacks by proposing Proactive Privacy Amnesia (PPA), which eliminates phone number exposure by 100% and reduces physical address exposure by 9.8% to 87.6% while maintaining model utility.
With the rise of large language models (LLMs), increasing research has recognized their risk of leaking personally identifiable information (PII) under malicious attacks. Although efforts have been made to protect PII in LLMs, existing methods struggle to balance privacy protection with maintaining model utility. In this paper, inspired by studies of amnesia in cognitive science, we propose a novel approach, Proactive Privacy Amnesia (PPA), to safeguard PII in LLMs while preserving their utility. This mechanism works by actively identifying and forgetting key memories most closely associated with PII in sequences, followed by a memory implanting using suitable substitute memories to maintain the LLM's functionality. We conduct evaluations across multiple models to protect common PII, such as phone numbers and physical addresses, against prevalent PII-targeted attacks, demonstrating the superiority of our method compared with other existing defensive techniques. The results show that our PPA method completely eliminates the risk of phone number exposure by 100% and significantly reduces the risk of physical address exposure by 9.8% - 87.6%, all while maintaining comparable model utility performance.