LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails

TL;DR

LPG introduces a dynamic safety guardrail for AI systems that balances reasoning and efficiency by compressing policy deliberation into latent states, enabling fast and accurate safety judgments.

Contribution

The paper presents LPG, a novel framework that learns semantic latent deliberation for dynamic safety policies, improving speed and accuracy without retraining.

Findings

LPG-4B achieves 84.5% safety accuracy and 77.9% F1 score.

LPG runs approximately 11 times faster than Qwen3-4B-Thinking.

LPG outperforms the strongest dynamic baseline on policy guardrail benchmarks.

Abstract

Guardrails are a critical safety layer for modern AI systems, but their operating regime is changing. As LLMs are deployed as customized assistants, safety policies are increasingly specified at inference time by users, organizations, or regulatory contexts. This makes safety enforcement fundamentally dynamic: the guardrail should adapt to changing safety policies without retraining. Yet this requirement creates a fundamental tension: faithfully judging complex policy contexts demands reasoning capability, while practical deployment requires low-latency responses. We introduce Latent Policy Guardrail (LPG), a guardrail framework that learnssemantic latent deliberation over dynamic policies. LPG compresses the internal deliberation needed for intent interpretation and policy grounding into continuous states supervised by decision-relevant semantics. At inference time, it generates only a compact verdict anchored to the violated policy clauses, preserving auditability while avoiding the latency of explicit reasoning. Across policy guardrail benchmarks, LPG-4B reaches 84.5% average safety accuracy and 77.9% F1 by compressing deliberation into just 10 latent tokens, outperforming the strongest dynamic baseline while running roughly 11 times faster than Qwen3-4B-Thinking under the single-sample evaluation setup. Code and data are available at https://github.com/SaFo-Lab/Latent_Policy_Guard.