{
  "id": 3459,
  "url": "https://arxiv.org/abs/2605.30808v1",
  "title": "Differentially Private Preference Data Synthesis for Large Language Model Alignment",
  "summary": "Preference alignment is a crucial post-training step for large language models (LLMs) to ensure their outputs align with human values. However, post-training on real human preference data raises privacy concerns, as these datasets often contain sensitive user prompts and human judgments. To address this, we propose DPPrefSyn, a novel algorithm for generating differentially private (DP) synthetic preference data to enable privacy-preserving preference alignment. DPPrefSyn is a principled framewor",
  "authors": "Fengyu Gao, Jing Yang",
  "category": "research",
  "topics": "safety-alignment,privacy-surveillance",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-29T03:53:12.000Z",
  "fetched_at": "2026-07-14T16:30:14.372Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/3459",
  "original_url": "https://arxiv.org/abs/2605.30808v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}