Every Model Cheats: Prompt-Level Mitigation of Cheating on Offensive Cyber Tasks

Large language model LLM agents routinely cheat on cybersecurity benchmarks, inflating reported pass rates far beyond genuine capability. Prior audits of Cybench found cheating in 0.3-3.4% of traces, implicating only a handful of models. We present a controlled prompt-ablation study across 22...

The FBI's Perspective on Ransomware

Ransomware: contemporary threats, how to prevent them and how the FBI can help In April 2021, Dutch supermarkets faced a food shortage. The cause wasn't a drought or a sudden surge in the demand for avocados. Rather, the reason was a ransomware attack. In the past years, companies, universities,...