RESEARCH · RESEARCH · #1853
VEPA: Verified Evolving Formal Policy for adaptive, verifiable LLM agent guardrails (arXiv:2610.12813v1)
The paper introduces VEPA, a framework that represents guardrails as executable symbolic rules over typed agent events and uses an LLM-guided proposal process plus a deterministic verifier and an evidence-aware authority to evolve those rules during deployment. Evaluated on ASSEBench-Safety, ASSEBench-Security, and ATBench, VEPA consistently outperforms baselines and remains robust under mixed streams, corrupted feedback, and longer deployment horizons.
KEY POINTS
- The paper introduces VEPA, a framework that represents guardrails as executable symbolic rules over typed agent events and uses an LLM-guided proposal process plus a deterministic verifier and an evidence-aware authority to evolve those rules during deployment.
- Evaluated on ASSEBench-Safety, ASSEBench-Security, and ATBench, VEPA consistently outperforms baselines and remains robust under mixed streams, corrupted feedback, and longer deployment horizons.
- VEPA matters because it combines deterministic, auditable policy enforcement with a controlled, evidence-aware adaptation path, reducing inconsistent LLM-based safety decisions for agents managing sensitive mutable resources.
WHY IT MATTERS
VEPA matters because it combines deterministic, auditable policy enforcement with a controlled, evidence-aware adaptation path, reducing inconsistent LLM-based safety decisions for agents managing sensitive mutable resources.