Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages
Resumo indisponível. Consulte o paper original.
PAPER / ARXIV:2610.12813
Duy Nguyen, Lesly Miculicich, Bhavana Dalvi Mishra, Shuo Li, Krishnamurthy Dj Dvijotham, Mohit Bansal, Tomas Pfister, Long T. Le
RESUMO
LLM agents increasingly operate over sensitive resources, such as private documents and mutable system states, which can create serious safety and security risks. Runtime guardrails should make inspectable, reproducible, and deterministic decisions before potentially irreversible agent actions. We introduce VEPA, a framework for the verified evolution of formal agent guardrails. VEPA represents the guardrail as a formal policy of executable symbolic rules, uses an LLM guided by deployment feedback to propose policy patches fixing newly observed failures while preserving previously correct behavior, then runs a deterministic verifier that only accepts structurally valid and conflict-free policies, plus an evidence-aware authority layer that activates a verified rule only once it has enough supporting evidence. VEPA consistently outperforms baselines across three benchmarks, including under mixed streams, corrupted feedback, and longer horizons.
NO MESMO MAPA
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.