How to DP-fy Your Data: A Practical Guide to Generating Synthetic Data With Differential Privacy · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2512.03238/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTHow to DP-fy Your Data: A Practical Guide to Generating Synthetic Data With Differential Privacy — mentioned by cududahttps://news.ycombinator.com/item?id=49680424urn:hackerpapers:mention:2512.03238:49680424Sun, 13 Sep 2026 05:46:53 GMT<p>Yep. It occurred to me a few days ago that the OAI TOS only says “Input” as what you provide “Output” as what you receive, Input and Output collectively as “Content.” and they won&#x27;t &quot;train&quot; on your content. But they can retain content for safety evaluations and debugging. What&#x27;s neat about &quot;safety evaluations&quot; in LLM parlance is apparently encountering any novel information constitutes a &quot;safety event&quot; that can result in new reinforcement learning data... This anthropic 2022 paper that basically describes how it&#x27;s a perpetual information siphoning machine with a cute little graphic…</p><p><a href="https://hackerpaper.kaipfister.de/p/2512.03238/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CR