# Keep It Private: Unsupervised Privatization of Online Text

arXiv: https://arxiv.org/abs/2405.10260 · HTML: https://mlsanity.com/paper/2405.10260/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 56 references that are in the map.
- First posted on arXiv 16 May 2024.

## Builds on (most cited first, 26 of 26)

- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [BLOOM: A 176B-Parameter Open-Access Multilingual Language Model](https://mlsanity.com/paper/2211.05100/) (2022)
- [Multilingual Denoising Pre-training for Neural Machine Translation](https://mlsanity.com/paper/2001.08210/) (2020)
- [Neural Network Acceptability Judgments](https://mlsanity.com/paper/1805.12471/) (2018)
- [Large Batch Optimization for Deep Learning: Training BERT in 76 minutes](https://mlsanity.com/paper/1904.00962/) (2019)
- [Self-critical Sequence Training for Image Captioning](https://mlsanity.com/paper/1612.00563/) (2016)
- [Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense](https://mlsanity.com/paper/2303.13408/) (2023)
- [PAWS: Paraphrase Adversaries from Word Scrambling](https://mlsanity.com/paper/1904.01130/) (2019)
- [The Pushshift Reddit Dataset](https://mlsanity.com/paper/2001.08435/) (2020)
- [Adversarial Example Generation with Syntactically Controlled Paraphrase Networks](https://mlsanity.com/paper/1804.06059/) (2018)
- [Addressing Semantic Drift in Question Generation for Semi-Supervised Question Answering](https://mlsanity.com/paper/1909.06356/) (2019)
- [Deep Communicating Agents for Abstractive Summarization](https://mlsanity.com/paper/1803.10357/) (2018)
- [Keep it Simple: Unsupervised Simplification of Multi-Paragraph Text](https://mlsanity.com/paper/2107.03444/) (2021)
- [Exploring Document-Level Literary Machine Translation with Parallel Paragraphs from World Literature](https://mlsanity.com/paper/2210.14250/) (2022)
- [A4NT: Author Attribute Anonymity by Adversarial Training of Neural Machine Translation](https://mlsanity.com/paper/1711.01921/) (2017)
- [Attribution and Obfuscation of Neural Text Authorship: A Data Mining Perspective](https://mlsanity.com/paper/2210.10488/) (2022)
- [ER-AE: Differentially Private Text Generation for Authorship Anonymization](https://mlsanity.com/paper/1907.08736/) (2019)
- [The Summary Loop: Learning to Write Abstractive Summaries Without Examples](https://mlsanity.com/paper/2105.05361/) (2021)
- [A Reinforced Topic-Aware Convolutional Sequence-to-Sequence Model for Abstractive Text Summarization](https://mlsanity.com/paper/1805.03616/) (2018)
- [Learning Invariant Representations of Social Media Users](https://mlsanity.com/paper/1910.04979/) (2019)
- [The Case for Being Average: A Mediocrity Approach to Style Masking and Author Obfuscation](https://mlsanity.com/paper/1707.03736/) (2017)
- [STEER: Unified Style Transfer with Expert Reinforcement](https://mlsanity.com/paper/2311.07167/) (2023)
- [Style Pooling: Automatic Text Style Obfuscation for Improved Classification Fairness](https://mlsanity.com/paper/2109.04624/) (2021)
- [Style Obfuscation by Invariance](https://mlsanity.com/paper/1805.07143/) (2018)

## Built on by (most cited first, 10 of 10)

- [Leveraging Multilingual Training for Authorship Representation: Enhancing Generalization across Languages and Domains](https://mlsanity.com/paper/2509.16531/) (2025)
- [Masks and Mimicry: Strategic Obfuscation and Impersonation Attacks on Authorship Verification](https://mlsanity.com/paper/2503.19099/) (2025)
- [Steering Large Language Models with Register Analysis for Arbitrary Style Transfer](https://mlsanity.com/paper/2505.00679/) (2025)
- [Can Large Language Models Really Recognize Your Name?](https://mlsanity.com/paper/2505.14549/) (2025)
- [Tau-Eval: A Unified Evaluation Framework for Useful and Private Text Anonymization](https://mlsanity.com/paper/2506.05979/) (2025)
- [Content Anonymization for Privacy in Long-form Audio](https://mlsanity.com/paper/2510.12780/) (2025)
- [Protecting De-identified Documents from Search-based Linkage Attacks](https://mlsanity.com/paper/2510.06383/) (2025)
- [Introducing the Privacy-HSD Trade-off: Hate Speech Detection, but not at the Cost of Privacy](https://mlsanity.com/paper/2608.19006/) (2026)
- [Aggregation-Aware Synthetic Text Generation Against Authorship Re-Identification](https://mlsanity.com/paper/2608.22161/) (2026)
- [PrivBench: A Holistic and Modular Benchmarking Platform for Evaluating Text-to-Text Privatization](https://mlsanity.com/paper/2608.29624/) (2026)

## Abstract (arXiv)

Authorship obfuscation techniques hold the promise of helping people protect their privacy in online communications by automatically rewriting text to hide the identity of the original author. However, obfuscation has been evaluated in narrow settings in the NLP literature and has primarily been addressed with superficial edit operations that can lead to unnatural outputs. In this work, we introduce an automatic text privatization framework that fine-tunes a large language model via reinforcement learning to produce rewrites that balance soundness, sense, and privacy. We evaluate it extensively on a large-scale test set of English Reddit posts by 68k authors composed of short-medium length texts. We study how the performance changes among evaluative conditions including authorial profile length and authorship detection strategy. Our method maintains high text quality according to both automated metrics and human evaluation, and successfully evades several automated authorship attacks.