FAVoR: Measuring and Mitigating Author-Style Homogenization in Federated Personalized Generation

2026-09-25

AI Summary

This study addresses the homogenization failure in federated parameter-efficient fine-tuning (PEFT), where distinct authorial styles tend to converge. To mitigate this issue, we propose FAVoR, a novel framework that incorporates a shared-private adapter architecture and introduces an author-style residual mechanism. This design effectively preserves personalized writing characteristics while strictly safeguarding data privacy. Experimental results demonstrate that FAVoR significantly enhances the recognizability of authorial style in generated texts, incurring only minimal degradation in semantic utility. Furthermore, external validation confirms the overall effectiveness of the proposed approach, highlighting its potential for privacy-preserving, personalized text generation within federated learning settings.

Abstract

Large language models are increasingly used as personalized writing assistants, but adapting a model across many authors can compromise individual writing style by pulling author-specific signals toward a shared register. Federated parameter-efficient fine-tuning (PEFT) offers a data-local setting for this multi-author adaptation problem: clients keep author text local while sharing compact adapter updates. However, we show that standard aggregation can preserve continuation utility while making different authors' generations less distinguishable in style space, a failure mode we define as author-style homogenization. We evaluate author-style retention with Angular Style Classification Encoder (ASCE)-based diagnostics on our main BlogText benchmark and ASCE-independent external authorship verification. Using this protocol, we find that common federated PEFT baselines can preserve semantic utility while averaging out author-specific signals. To address this homogenization, we instantiate FAVoR (Federated Authorial Voice Retention), an author-style residual mechanism for federated PEFT. FAVoR uses a shared-private adapter design: clients upload shared-adapter updates while retaining author-specific residual corrections locally. Across BlogText and external Mythos-Reddit validation, FAVoR improves author-style retention over standard and personalized federated PEFT baselines. These gains come with small continuation-utility trade-offs and are supported by component ablations, external verification, and cold-start transfer.

Problem

• author-style homogenization
• federated PEFT
• personalized generation
• large language models
• style retention

Innovation

• Federated PEFT
• Author-style homogenization
• Shared-private adapter
• Style residual mechanism
• Personalized generation