Private Seeds, Public LLMs: Realistic and Privacy-Preserving Synthetic Data Generation
Qian Ma, Sarah Rajtmajer · 2026
Read Private Seeds, Public LLMs: Realistic and Privacy-Preserving Synthetic Data Generation by Qian Ma, Sarah Rajtmajer (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07486.txt.
Description
Ma and Rajtmajer propose RPSG, a method for making realistic synthetic copies of private text while keeping privacy. Large language models excel at synthetic data, but private-text replication demands a careful balance of privacy and utility. RPSG starts from private seeds and uses privacy-preserving tactics, including a formal differential-privacy rule in picking outputs. Broad tests against leading private synthetic-data methods show RPSG achieves high closeness to private data while providing strong safety.
Catalog record
| CREATOR | Qian Ma, Sarah Rajtmajer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 79319 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.07486