aiqbench analysis
SharedRequest offers a practical solution for protecting user prompt privacy in public LLM deployments by operating at the batch level without requiring model-specific modifications. By grouping semantically equivalent instructions to amortize inference costs, the framework achieves up to a fivefold reduction in query expenses while maintaining higher utility than existing differential privacy baselines. This approach is particularly valuable for organizations seeking to deploy privacy-preserving inference across diverse model architectures without incurring significant architectural overhead.
SharedRequest: Privacy-Preserving Model-Agnostic Inference for Large Language Models
arXiv:2606.05004v1 Announce Type: cross Abstract: With the widespread deployment of public large language models (LLMs) such as ChatGPT, protecting user prompt privacy has become an increasingly critical issue. Existing privacy-preserving inference methods sacrifice either utility or efficiency, and often require model-specific modifications that limit their compatibility. In this paper, we propose SharedRequest, a model-agnostic framework for privacy-preserving LLM inference that reformulates privacy protection at the batch level rather than the individual-prompt level. The key idea is to obscure sensitive information by mixing original prompts with noisy variants, while grouping semantically equivalent instructions to amortize the inference cost over a large batch of queries with minimal impact on LLM response quality. This design is independent of the LLM architecture, requiring no access to model parameters or architectural modification. Empirical results demonstrate that SharedRequest achieves over $20\%$ higher utility compared to prior differential privacy baselines, and its shared-prompt mechanism reduces query cost by up to $5\times$ compared to non-batched inference.
Abstract:With the widespread deployment of public large language models (LLMs) such as ChatGPT, protecting user prompt privacy has become an increasingly critical issue. Existing privacy-preserving inference methods sacrifice either utility or efficiency, and often require model-specific modifications that limit their compatibility. In this paper, we propose SharedRequest, a model-agnostic framework for privacy-preserving LLM inference that reformulates privacy protection at the batch level rather than the individual-prompt level. The key idea is to obscure sensitive information by mixing original prompts with noisy variants, while grouping semantically equivalent instructions to amortize the inference cost over a large batch of queries with minimal impact on LLM response quality. This design is independent of the LLM architecture, requiring no access to model parameters or architectural modification. Empirical results demonstrate that SharedRequest achieves over $20\%$ higher utility compared to prior differential privacy baselines, and its shared-prompt mechanism reduces query cost by up to $5\times$ compared to non-batched inference.
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
arXiv:2606.05004 [cs.CR]
(or arXiv:2606.05004v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2606.05004 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 251 words
Related coverage
[AINews] not much happened today
Latent Space · Sep 18, 2026
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · May 21, 2026
Privacy-Aware Decoding: Mitigating Privacy Leakage of Large Language Models in Retrieval-Augmented Generation
arXiv · Jun 3, 2026
QuanText: Protecting Dataset-Level Secrets in Textual Data Sharing
arXiv · Sep 17, 2026