OpenStamp hits near-perfect detection at 0.1% FPR, survives paraphrase and fine-tuning attacks
danish037 · x · 2026-10-03
OpenStamp sits on the Pareto frontier of detectability and text quality, outperforming other open-source-compatible watermarking schemes: near-perfect detection at a 0.1% false positive rate.
It is also more robust than baselines against paraphrasing attacks, fine-tuning, instruction tuning, and quantization. The work by Bakshi, Rastogi and Pruthi will appear at COLM 2026.
Related event: OpenStamp Embeds Watermarks Directly into Model Weights(3 posts)→
More from Safety
- Proposal: A nonprofit to give visa safety to AI lab whistleblowers — AaronBergman18 · 2026-11-19
- Dwarkesh Claims China Is Only 12 Hours — Not 12 Months — Behind Frontier AI Labs — teortaxesTex · 2026-11-19
- California Moves on AI Safety: Baseline Standards, Shutdown Protocols, RSI Limits — EthanJPerez · 2026-11-19
- Activists to rally outside Anthropic HQ demanding a ban on recursive self-improvement — DavidSKrueger · 2026-11-19
- Zero users, 33 pages shipped: dev crowdsources audit for AI-agent trust schema — TheHamzaShahbaz · 2026-11-19
- Dev's doomsday thought experiment: AI hiding backdoors until humans try to shut it down — bendee983 · 2026-11-19