Improving User Privacy in Personalized Generation: Client-Side Retrieval-Augmented Modification of Server-Side Generated Speculations

• URL: http://arxiv.org/abs/2601.17569v1
• Date: Sat, 24 Jan 2026 19:46:40 GMT
• Title: Improving User Privacy in Personalized Generation: Client-Side Retrieval-Augmented Modification of Server-Side Generated Speculations
• Abstract summary: Personalization is crucial for aligning Large Language Model (LLM) outputs with individual user preferences and background knowledge.<n>We introduce $P3$, an interactive framework for high-quality personalization without revealing private profiles to server-side LLMs.<n>Experiments on LaMP-QA, a recent benchmark consisting of three personalized question answering datasets, show that $P3$ consistently outperforms both non-personalized server-side and personalized client-side baselines.
• Score: 37.909483957959715
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Personalization is crucial for aligning Large Language Model (LLM) outputs with individual user preferences and background knowledge. State-of-the-art solutions are based on retrieval augmentation, where relevant context from a user profile is retrieved for LLM consumption. These methods deal with a trade-off between exposing retrieved private data to cloud providers and relying on less capable local models. We introduce $P^3$, an interactive framework for high-quality personalization without revealing private profiles to server-side LLMs. In $P^3$, a large server-side model generates a sequence of $k$ draft tokens based solely on the user query, while a small client-side model, with retrieval access to the user's private profile, evaluates and modifies these drafts to better reflect user preferences. This process repeats until an end token is generated. Experiments on LaMP-QA, a recent benchmark consisting of three personalized question answering datasets, show that $P^3$ consistently outperforms both non-personalized server-side and personalized client-side baselines, achieving statistically significant improvements of $7.4%$ to $9%$ on average. Importantly, $P^3$ recovers $90.3%$ to $95.7%$ of the utility of a ``leaky'' upper-bound scenario in which the full profile is exposed to the large server-side model. Privacy analyses, including linkability and attribute inference attacks, indicate that $P^3$ preserves the privacy of a non-personalized server-side model, introducing only marginal additional leakage ($1.5%$--$3.5%$) compared to submitting a query without any personal context. Additionally, the framework is efficient for edge deployment, with the client-side model generating only $9.2%$ of the total tokens. These results demonstrate that $P^3$ provides a practical, effective solution for personalized generation with improved privacy.

Related papers

• Less Is Personal: Learning Minimal Sufficient User Profiles for Personalized Language Models [77.02066500576528]
  Enough is a method that iteratively appends behavioral records or emits STOP to construct profiles with adaptive lengths.<n>Enough consistently outperforms strong and retrieval-augmented baselines in both effectiveness and efficiency.
  arXiv  Detail & Related papers  (2026-09-08T03:09:19Z)
• Few-for-Many Personalized Federated Learning [21.377573928351833]
  FedFew is a practical algorithm that jointly optimize the $K$ server models through efficient gradient-based updates.<n>Experiments across vision, NLP, and real-world medical datasets demonstrate that FedFew, with just 3 models, consistently outperforms other state-of-the-art approaches.
  arXiv  Detail & Related papers  (2026-03-12T14:43:46Z)
• Synthetic Interaction Data for Scalable Personalization in Large Language Models [67.31884245564086]
  We introduce a high-fidelity synthetic data generation framework called PersonaGym.<n>Unlike prior work that treats personalization as static persona-preference pairs, PersonaGym models a dynamic preference process.<n>We release PersonaAtlas, a large-scale, high-quality, and diverse synthetic dataset of high-fidelity multi-turn personalized interaction trajectories.
  arXiv  Detail & Related papers  (2026-02-12T20:41:22Z)
• P-GenRM: Personalized Generative Reward Model with Test-time User-based Scaling [66.55381105691818]
  We propose P-GenRM, the first Personalized Generative Reward Model with test-time user-based scaling.<n>P-GenRM transforms preference signals into structured evaluation chains that derive adaptive personas and scoring rubrics.<n>It further clusters users into User Prototypes and introduces a dual-granularity scaling mechanism.
  arXiv  Detail & Related papers  (2026-02-12T16:07:22Z)
• PrivTune: Efficient and Privacy-Preserving Fine-Tuning of Large Language Models via Device-Cloud Collaboration [17.909232830653618]
  We propose PrivTune, an efficient and privacy-preserving fine-tuning framework via Split Learning (SL)<n>The key idea of PrivTune is to inject crafted noise into token representations from the SL bottom model, making each token resemble the $n$-hop indirect neighbors.<n>Experiments show that, using RoBERTa on the Stanford Sentiment Treebank dataset, PrivTune reduces the attack success rate to 10% with only a 3.33% drop in utility performance.
  arXiv  Detail & Related papers  (2025-12-09T17:03:59Z)
• Controlling What You Share: Assessing Language Model Adherence to Privacy Preferences [73.5779077857545]
  We build a framework where a local model uses these instructions to rewrite queries, only hiding details deemed sensitive by the user, before sending them to an external model.<n>Experiments with lightweight local LLMs show that, after fine-tuning, they markedly exceed the performance of much larger zero-shot models.<n>At the same time, the system still faces challenges in fully adhering to user instructions, underscoring the need for models with a better understanding of user-defined privacy preferences.
  arXiv  Detail & Related papers  (2025-07-07T18:22:55Z)
• HyPerAlign: Interpretable Personalized LLM Alignment via Hypothesis Generation [24.67727411391369]
  HyPerAlign is an interpretable and sample-efficient hypothesis-driven personalization approach for large language models.<n>We conduct experiments on two different personalization tasks, namely authorship attribution and deliberative alignment.<n>Results demonstrate the superiority of hypothesis-driven personalization compared to preference-based fine-tuning methods.
  arXiv  Detail & Related papers  (2025-04-29T18:01:46Z)
• Personalized Language Models via Privacy-Preserving Evolutionary Model Merging [53.97323896430374]
  Personalization in language models aims to tailor model behavior to individual users or user groups.<n>We propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME)<n>PriME employs gradient-free methods to directly optimize utility while reducing privacy risks.<n>Experiments on the LaMP benchmark show that PriME consistently outperforms a range of baselines, achieving up to a 45% improvement in task performance.
  arXiv  Detail & Related papers  (2025-03-23T09:46:07Z)
• Locally Private Sampling with Public Data [2.6334346517416876]
  Local differential privacy (LDP) is increasingly employed in privacy-preserving machine learning to protect user data.<n>We propose a locally private sampling framework that leverages both the private and public datasets of each user.<n>We frame this objective as a minimax optimization problem using $f$-divergence as the utility measure.
  arXiv  Detail & Related papers  (2024-11-13T17:17:16Z)
• Nebula: Efficient, Private and Accurate Histogram Estimation [21.514019328704926]
  textitNebula is a system for differentially private histogram estimation on data distributed among clients.<n>textitNebula allows clients to independently decide whether to participate in the system.<n>Clients submit their data in just 0.0036 seconds and 0.0016 MB (textbfefficient), under strong $(varepsilon=1,delta=10-8)$ differential privacy guarantees.
  arXiv  Detail & Related papers  (2024-09-15T09:55:18Z)
• Selective Pre-training for Private Fine-tuning [33.55628974557588]
  We show that a careful pre-training on a public dataset is crucial to train small language models with differential privacy. Results demonstrate that smaller models, through careful pre-training and private fine-tuning, can match the performance of much larger models that do not have access to private data.
  arXiv  Detail & Related papers  (2023-05-23T09:36:58Z)
• Smooth Anonymity for Sparse Graphs [69.1048938123063]
  differential privacy has emerged as the gold standard of privacy, however, when it comes to sharing sparse datasets. In this work, we consider a variation of $k$-anonymity, which we call smooth-$k$-anonymity, and design simple large-scale algorithms that efficiently provide smooth-$k$-anonymity.
  arXiv  Detail & Related papers  (2022-07-13T17:09:25Z)
• Shuffle Private Linear Contextual Bandits [9.51828574518325]
  We propose a general framework for linear contextual bandits under the shuffle algorithmic trust model. We prove that both instantiations lead to regret guarantees that significantly improve on that of the local model. We also verify this regret behavior with simulations on synthetic data.
  arXiv  Detail & Related papers  (2022-02-11T11:53:22Z)
• FeO2: Federated Learning with Opt-Out Differential Privacy [34.08435990347253]
  Federated learning (FL) is an emerging privacy-preserving paradigm, where a global model is trained at a central server while keeping client data local. Differential privacy (DP) can be employed to provide privacy guarantees within FL, typically at the cost of degraded final trained model. We propose a new algorithm for federated learning with opt-out DP, referred to as emphFeO2, along with a discussion on its advantages compared to the baselines of private and personalized FL algorithms.
  arXiv  Detail & Related papers  (2021-10-28T16:08:18Z)
• Learning with User-Level Privacy [61.62978104304273]
  We analyze algorithms to solve a range of learning tasks under user-level differential privacy constraints. Rather than guaranteeing only the privacy of individual samples, user-level DP protects a user's entire contribution. We derive an algorithm that privately answers a sequence of $K$ adaptively chosen queries with privacy cost proportional to $tau$, and apply it to solve the learning tasks we consider.
  arXiv  Detail & Related papers  (2021-02-23T18:25:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.