# RL-Finetuned LLMs for Privacy-Preserving Synthetic Rewriting

arXiv: https://arxiv.org/abs/2508.19286 · HTML: https://mlsanity.com/paper/2508.19286/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 38 references that are in the map.
- First posted on arXiv 25 Aug 2025.

## Builds on (most cited first, 30 of 30)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Deep Learning with Differential Privacy](https://mlsanity.com/paper/1607.00133/) (2016)
- [Extracting Training Data from Large Language Models](https://mlsanity.com/paper/2012.07805/) (2020)
- [ERASER: A Benchmark to Evaluate Rationalized NLP Models](https://mlsanity.com/paper/1911.03429/) (2019)
- [Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense](https://mlsanity.com/paper/2303.13408/) (2023)
- [Black-box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers](https://mlsanity.com/paper/1801.04354/) (2018)
- [Analyzing Leakage of Personally Identifiable Information in Language Models](https://mlsanity.com/paper/2302.00539/) (2023)
- [Beyond Memorization: Violating Privacy Via Inference with Large Language Models](https://mlsanity.com/paper/2310.07298/) (2023)
- [Same Author or Just Same Topic? Towards Content-Independent Style Representations](https://mlsanity.com/paper/2204.04907/) (2022)
- [Differentially Private Language Models for Secure Data Sharing](https://mlsanity.com/paper/2210.13918/) (2022)
- [Locally Differentially Private Document Generation Using Zero Shot Prompting](https://mlsanity.com/paper/2310.16111/) (2023)
- [Generalised Differential Privacy for Text Document Processing](https://mlsanity.com/paper/1811.10256/) (2018)
- [A Synthetic Dataset for Personal Attribute Inference](https://mlsanity.com/paper/2406.07217/) (2024)
- [StyleDistance: Stronger Content-Independent Style Embeddings with Synthetic Parallel Examples](https://mlsanity.com/paper/2410.12757/) (2024)
- [Privacy Regularization: Joint Privacy-Utility Optimization in Language Models](https://mlsanity.com/paper/2103.07567/) (2021)
- [DP-BART for Privatized Text Rewriting under Local Differential Privacy](https://mlsanity.com/paper/2302.07636/) (2023)
- [Adversarial Robustness of Neural-Statistical Features in Detection of Generative Transformers](https://mlsanity.com/paper/2203.07983/) (2022)
- [DP-MLM: Differentially Private Text Rewriting Using Masked Language Models](https://mlsanity.com/paper/2407.00637/) (2024)
- [Measuring Data Diversity for Instruction Tuning: A Systematic Analysis and A Reliable Metric](https://mlsanity.com/paper/2502.17184/) (2025)
- [StyleRemix: Interpretable Authorship Obfuscation via Distillation and Perturbation of Style Elements](https://mlsanity.com/paper/2408.15666/) (2024)
- [The Case for Being Average: A Mediocrity Approach to Style Masking and Author Obfuscation](https://mlsanity.com/paper/1707.03736/) (2017)
- [R.R.: Unveiling LLM Training Privacy through Recollection and Ranking](https://mlsanity.com/paper/2502.12658/) (2025)
- [Just Rewrite It Again: A Post-Processing Method for Enhanced Semantic Similarity and Privacy Preservation of Differentially Private Rewritten Text](https://mlsanity.com/paper/2405.19831/) (2024)
- [A Multi-Faceted Evaluation Framework for Assessing Synthetic Data Generated by Large Language Models](https://mlsanity.com/paper/2404.14445/) (2024)
- [Casper: Prompt Sanitization for Protecting User Privacy in Web-Based Large Language Models](https://mlsanity.com/paper/2408.07004/) (2024)
- [ALISON: Fast and Effective Stylometric Authorship Obfuscation](https://mlsanity.com/paper/2402.00835/) (2024)
- [Granularity is crucial when applying differential privacy to text: An investigation for neural machine translation](https://mlsanity.com/paper/2407.18789/) (2024)

## Abstract (arXiv)

The performance of modern machine learning systems depends on access to large, high-quality datasets, often sourced from user-generated content or proprietary, domain-specific corpora. However, these rich datasets inherently contain sensitive personal information, raising significant concerns about privacy, data security, and compliance with regulatory frameworks. While conventional anonymization techniques can remove explicit identifiers, such removal may result in performance drop in downstream machine learning tasks. More importantly, simple anonymization may not be effective against inference attacks that exploit implicit signals such as writing style, topical focus, or demographic cues, highlighting the need for more robust privacy safeguards during model training. To address the challenging issue of balancing user privacy and data utility, we propose a reinforcement learning framework that fine-tunes a large language model (LLM) using a composite reward function that jointly optimizes for explicit and implicit privacy, semantic fidelity, and output diversity. To effectively capture population level regularities, the privacy reward combines semantic cues with structural patterns derived from a minimum spanning tree (MST) over latent representations. By modeling these privacy-sensitive signals in their distributional context, the proposed approach guides the model to generate synthetic rewrites that preserve utility while mitigating privacy risks. Empirical results show that the proposed method significantly enhances author obfuscation and privacy metrics without degrading semantic quality, providing a scalable and model-agnostic solution for privacy preserving data generation in the era of large language models.