Metraipolis › archive › dc_7151f8836091

Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Models — arXiv:2609.32706

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32706
License (as archived)
CC0-1.0
Archived
2026-10-07T21:00:33.607Z
SHA-256 of the archived text
fec8d842dacc02b5465746834990e9cd95c6907d346ecd807d693ae2224bb24c
Stable URL of this page
https://metraipolis.com/documento/dc_7151f8836091

How to cite

Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Models — arXiv:2609.32706 — arXiv (en), https://arxiv.org/abs/2609.32706 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7151f8836091 (Metraipolis archive dc_7151f8836091, sha256 fec8d842dacc02b5465746834990e9cd95c6907d346ecd807d693ae2224bb24c)
BibTeX
@misc{dc7151f8836091,
  title = {Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Models — arXiv:2609.32706},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32706},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7151f8836091; sha256 fec8d842dacc02b5465746834990e9cd95c6907d346ecd807d693ae2224bb24c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32706. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Cloud-based large language models (LLMs) require users to disclose plaintext data to service providers, creating privacy risks in sensitive domains. Existing privacy-preserving approaches often trade utility for protection, incur substantial computational or communication overhead, remain vulnerable to reconstruction from intermediate representations, or protect only a subset of the training and inference pipeline. We introduce Client-Resolved Generation (CRG), a genera- tion interface that separates server-side generation from the lexical realization of input-derived content. The client transmits only pooled and noise-perturbed rep- resentations, while input-derived output content is represented using request-local positional references and resolved to its original strings only on the client. This interface protects private input and input-derived output content during both train- ing and inference while allowing the service provider to keep its proprietary model parameters hidden from the client. At the same time, exact lexical reuse remains possible without directly exposing the reused content on the provider-visible gen- eration path. We evaluate CRG on medical and document-grounded QA, sensi- tive identifier transfer, and tool calling, together with reconstruction and raw-logit leakage analyses. On SealTools, CRG improves complete-call exact match from 57.3% to 79.9% over the input-privacy framework PPFT, with larger gains as more required output content can be resolved through references. Together, these results show that CRG provides a practical interface for privacy-sensitive cloud LLMs by reducing plaintext exposure across both input and output pathways while preserv- ing task utility and server-side model confidentiality.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.32706 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T20:29:02.440Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).