R EAL G UARD: Reliable, Efficient, and Auditable LLM Safeguards via Latent Reasoning
cs.AI
Submitted: 2026-08-04
Updated: 2026-10-13
Terminology
Sources
- Reasoning Models Don't Always Say What They Think
- Compressed Chain of Thought: Efficient Reasoning Through Dense Representations
- Implicit Chain of Thought Reasoning via Knowledge Distillation
- AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts
- Training Large Language Models to Reason in a Continuous Latent Space
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety
- Measuring Faithfulness in Chain-of-Thought Reasoning
- GSPR: Aligning LLM Safeguards as Generalizable Safety Policy Reasoners
- GuardReasoner: Towards Reasoning-based LLM Safeguards
- The Llama 3 Herd of Models
- Taxonomy-Adaptive Moderation Model with Robust Guardrails for Large Language Models
- SimpleSafetyTests: a Test Suite for Identifying Critical Safety Risks in Large Language Models
- OmniGuard: Unified Omni-Modal Guardrails with Deliberate Reasoning
- GuardReasoner-Omni: A Reasoning-based Multi-modal Guardrail for Text, Image, Video, and Audio
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems