Human research and writing on AI security, based on our own experiments.Human research and writing on AI security, based on our own experiments.Forschung und Texte von Menschen zur KI-Sicherheit, auf Grundlage unserer eigenen Experimente.
What happens to security when you distract and confuse an agent?
Natalie Ma
Notes from replicating LLMLingua-2 + why compression is a safety issue