You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Official AISTATS 2026 implementation formalizing Sensitivity Awareness (SA) via Differential Privacy. Features a 4-bit LoRA post-training pipeline to secure LLM agents against contextual authorization and access-control bypasses.
The first benchmark environment for Sensitivity Awareness (SA) in LLMs. Evaluating how language model agents handle Role-Based Access Control (RBAC), Confused Deputy vulnerabilities, and Contextual Authorization rules.