Yesterday's Shield, Today's Spear: A Self-Evolving Safety Guardrail in Production
Cong Ming, Jingyi Chen, Bin Liu, Qi Chu, Tao Gong, Nenghai Yu, Yingfei Xiang
cs.AI
Submitted: 2026-08-09
Updated: 2026-08-11
Code: https://github.com/Trams1017/SESG
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Constitutional AI: Harmlessness from AI Feedback
- On the Opportunities and Risks of Foundation Models
- Lattice: Generative Guardrails for Conversational Agents
- M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
- Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense
- Building Guardrails for Large Language Models
- AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts
- Virus: Harmful Fine-tuning Attack for Large Language Models Bypassing Guardrail Moderation
- Obscure but Effective: Classical Chinese Jailbreak Prompt Optimization via Bio-Inspired Search
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- DeepInception: Hypnotize Large Language Model to Be Jailbreaker
- YuFeng-XGuard: A Reasoning-Centric, Interpretable, and Flexible Guardrail Model for Large Language Models
- SafeHarbor: Defining Precise Decision Boundaries via Hierarchical Memory-Augmented Guardrail for LLM Agent Safety
- CodeChameleon: Personalized Encryption Framework for Jailbreaking Large Language Models
- HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs
- Jailbreak Mimicry: Automated Discovery of Narrative-Based Jailbreaks for Large Language Models
- Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical Expression
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- OpenAI GPT-5 System Card
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