LLM Watermarking Using Mixtures and Statistical-To-Computational Gaps

Given a text, can we determine whether it was generated by a large language model LLM or by a human? A widely studied approach to this problem is watermarking. We propose an undetectable and elementary watermarking scheme in the closed setting. Also, in the harder open setting, where the adversar...

chromium-browser: Settings allowed executable files to run immediately after download in unsepcified

Google Chrome before 44.0.2403.89 does not ensure that the auto-open list omits all dangerous file types, which makes it easier for remote attackers to execute arbitrary code by providing a crafted file and leveraging a user's previous "Always open files of this type" choice, related to...