Linear Ensembles Wash Away Watermarks: On the Fragility of Distributional Perturbations in LLMs 事件

Linear Ensembles Wash Away Watermarks: On the Fragility of Distributional Perturbations in LLMs arXiv:2605.30501v1 Announce Type: new Abstract: Watermarking embeds statistical signatures in AI-generated text for detection and attribution. We reveal a fundamental vulnerability: when users access multiple models (today's reality), watermarks trivially fail. Watermarks perturb output distributions away from the original, and in competitive markets, these perturbations are typically independent acr