OSNIP: Breaking the Privacy-Utility-Efficiency Trilemma in LLM Inference via Obfuscated Semantic Null Space
This addresses privacy concerns for users of LLM inference services by providing a lightweight encryption method, though it appears incremental as it builds on geometric intuitions for privacy enhancement.
The paper tackled the challenge of balancing privacy, utility, and efficiency in LLM inference by proposing OSNIP, a client-side encryption framework that injects perturbations into an obfuscated semantic null space, resulting in state-of-the-art performance with sharply reduced attack success rates and strong model utility on 12 benchmarks.
We propose Obfuscated Semantic Null space Injection for Privacy (OSNIP), a lightweight client-side encryption framework for privacy-preserving LLM inference. Generalizing the geometric intuition of linear kernels to the high-dimensional latent space of LLMs, we formally define the ``Obfuscated Semantic Null Space'', a high-dimensional regime that preserves semantic fidelity while enforcing near-orthogonality to the original embedding. By injecting perturbations that project the original embedding into this space, OSNIP ensures privacy without any post-processing. Furthermore, OSNIP employs a key-dependent stochastic mapping that synthesizes individualized perturbation trajectories unique to each user. Evaluations on 12 generative and classification benchmarks show that OSNIP achieves state-of-the-art performance, sharply reducing attack success rates while maintaining strong model utility under strict security constraints.