Cover Image

Silencing the Guardrails: Inference-Time Jailbreaking via Dynamic Contextual Representation Ablation

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Xing, Wenpeng, Fang, Moran, Wang, Guangtai, Lin, Changting, Han, Meng
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

HGMF: A Hierarchical Gaussian Mixture Framework for Scalable Tool Invocation within the Model Context Protocol
by: Xing, Wenpeng, et al.
Published: (2025)

LLMs can be Dangerous Reasoners: Analyzing-based Jailbreak Attack on Large Language Models
by: Lin, Shi, et al.
Published: (2024)

NeuRel-Attack: Neuron Relearning for Safety Disalignment in Large Language Models
by: Zhou, Yi, et al.
Published: (2025)

CoDe-NeRF: Neural Rendering via Dynamic Coefficient Decomposition
by: Xing, Wenpeng, et al.
Published: (2025)

DIAP: A Decentralized Agent Identity Protocol with Zero-Knowledge Proofs and a Hybrid P2P Stack
by: Liu, Yuanjie, et al.
Published: (2025)

Latent Fusion Jailbreak: Blending Harmful and Harmless Representations to Elicit Unsafe LLM Outputs
by: Xing, Wenpeng, et al.
Published: (2025)

CTCC: A Robust and Stealthy Fingerprinting Framework for Large Language Models via Cross-Turn Contextual Correlation Backdoor
by: Xu, Zhenhua, et al.
Published: (2025)

SproutBench: A Benchmark for Safe and Ethical Large Language Models for Youth
by: Xing, Wenpeng, et al.
Published: (2025)

LatentAudit: Real-Time White-Box Faithfulness Monitoring for Retrieval-Augmented Generation with Verifiable Deployment
by: Yu, Zhe, et al.
Published: (2026)

UW-3DGS: Underwater 3D Reconstruction with Physics-Aware Gaussian Splatting
by: Xing, Wenpeng, et al.
Published: (2025)

From Retinal Evidence to Safe Decisions: RETINA-SAFE and ECRT for Hallucination Risk Triage in Medical LLMs
by: Yu, Zhe, et al.
Published: (2026)

MCP-Guard: A Multi-Stage Defense-in-Depth Framework for Securing Model Context Protocol in Agentic AI
by: Xing, Wenpeng, et al.
Published: (2025)

MIST: Jailbreaking Black-box Large Language Models via Iterative Semantic Tuning
by: Zheng, Muyang, et al.
Published: (2025)

SoK: Evaluating Jailbreak Guardrails for Large Language Models
by: Wang, Xunguang, et al.
Published: (2025)

Guarding the Guardrails: A Taxonomy-Driven Approach to Jailbreak Detection
by: Giarrusso, Francesco, et al.
Published: (2025)

Bridging the Safety Gap: A Guardrail Pipeline for Trustworthy LLM Inferences
by: Han, Shanshan, et al.
Published: (2025)

Active Honeypot Guardrail System: Probing and Confirming Multi-Turn LLM Jailbreaks
by: Wu, ChenYu, et al.
Published: (2025)

Provably Secure Agent Guardrail
by: Wu, Benlong, et al.
Published: (2026)

ConceptGuard: Neuro-Symbolic Safety Guardrails via Sparse Interpretable Jailbreak Concepts
by: Aswal, Darpan, et al.
Published: (2025)

Towards Robust and Secure Embodied AI: A Survey on Vulnerabilities and Attacks
by: Xing, Wenpeng, et al.
Published: (2025)

MO-RiskVAE: A Multi-Omics Variational Autoencoder for Survival Risk Modeling in Multiple MyelomaMO-RiskVAE
by: Chen, Zixuan, et al.
Published: (2026)

ForgetMark: Stealthy Fingerprint Embedding via Targeted Unlearning in Language Models
by: Xu, Zhenhua, et al.
Published: (2026)

Cordon-MAS: Defending RAG against Knowledge Poisoning via Information-Flow Control
by: Yu, Zhe, et al.
Published: (2026)

TriLens: Per-Layer Logit-Lens Entropy for White-Box Hallucination Detection
by: Yang, Bohan, et al.
Published: (2026)

Disentangled Safety Adapters Enable Efficient Guardrails and Flexible Inference-Time Alignment
by: Krishna, Kundan, et al.
Published: (2025)

PiCo: Jailbreaking Multimodal Large Language Models via Pictorial Code Contextualization
by: Liu, Aofan, et al.
Published: (2025)

"Moralized" Multi-Step Jailbreak Prompts: Black-Box Testing of Guardrails in Large Language Models for Verbal Attacks
by: Wang, Libo
Published: (2024)

Composition Collapse: Stable Factual Knowledge Does Not Imply Compositional Reasoning
by: Yu, Zhe, et al.
Published: (2026)

Spiral of Silence in Large Language Model Agents
by: Zhong, Mingze, et al.
Published: (2025)

DNF: Dual-Layer Nested Fingerprinting for Large Language Model Intellectual Property Protection
by: Xu, Zhenhua, et al.
Published: (2026)

Understanding and Defending VLM Jailbreaks via Jailbreak-Related Representation Shift
by: Wei, Zhihua, et al.
Published: (2026)

Learning What Matters Now: Dynamic Preference Inference under Contextual Shifts
by: Cao, Xianwei, et al.
Published: (2026)

UniGuard: Towards Universal Safety Guardrails for Jailbreak Attacks on Multimodal Large Language Models
by: Oh, Sejoon, et al.
Published: (2024)

Direct Behavior Optimization: Unlocking the Potential of Lightweight LLMs
by: Yang, Hongming, et al.
Published: (2025)

Efficient LLM-Jailbreaking via Multimodal-LLM Jailbreak
by: Ji, Haoxuan, et al.
Published: (2024)

Test-Time Training Undermines Safety Guardrails
by: Antonelli, Simone, et al.
Published: (2026)

The Attribution Blind Spot: Detecting When Language Models Rely on Memory Rather Than Retrieved Context
by: Yu, Zhe, et al.
Published: (2026)

Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment
by: Ghosal, Soumya Suvra, et al.
Published: (2024)

MEraser: An Effective Fingerprint Erasure Approach for Large Language Models
by: Zhang, Jingxuan, et al.
Published: (2025)

Spectral Logit Sculpting: Adaptive Low-Rank Logit Transformation for Controlled Text Generation
by: Li, Jin, et al.
Published: (2025)