Cover Image

Safety Context Injection: Inference-Time Safety Alignment via Static Filtering and Agentic Analysis

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Xu, Zhenhao, Chang, Wenhan, Chen, Yichuan, Fang, Yuxin, Liu, Junhao, Zhu, Tianqing
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

From Thinking to Output: Chain-of-Thought and Text Generation Characteristics in Reasoning Language Models
by: Liu, Junhao, et al.
Published: (2025)

Unreal Thinking: Chain-of-Thought Hijacking via Two-stage Backdoor
by: Chang, Wenhan, et al.
Published: (2026)

Zero-shot Class Unlearning via Layer-wise Relevance Analysis and Neuronal Path Perturbation
by: Chang, Wenhan, et al.
Published: (2024)

Layer-Aware Representation Filtering: Purifying Finetuning Data to Preserve LLM Safety Alignment
by: Li, Hao, et al.
Published: (2025)

Agent Safety Alignment via Reinforcement Learning
by: Sha, Zeyang, et al.
Published: (2025)

Large Language Models Merging for Enhancing the Link Stealing Attack on Graph Neural Networks
by: Guan, Faqian, et al.
Published: (2024)

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

Beyond Surface Alignment: Rebuilding LLMs Safety Mechanism via Probabilistically Ablating Refusal Direction
by: Xie, Yuanbo, et al.
Published: (2025)

Reimagining Safety Alignment with An Image
by: Xia, Yifan, et al.
Published: (2025)

Chain-of-Lure: A Universal Jailbreak Attack Framework using Unconstrained Synthetic Narratives
by: Chang, Wenhan, et al.
Published: (2025)

Involuntary In-Context Learning: Exploiting Few-Shot Pattern Completion to Bypass Safety Alignment in GPT-5.4
by: Polyakov, Alex, et al.
Published: (2026)

SafeThinker: Reasoning about Risk to Deepen Safety Beyond Shallow Alignment
by: Fang, Xianya, et al.
Published: (2026)

ReasAlign: Reasoning Enhanced Safety Alignment against Prompt Injection Attack
by: Li, Hao, et al.
Published: (2026)

From Threat to Tool: Leveraging Refusal-Aware Injection Attacks for Safety Alignment
by: Chae, Kyubyung, et al.
Published: (2025)

From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI
by: Zhang, Zelin, et al.
Published: (2026)

Really Unlearned? Verifying Machine Unlearning via Influential Sample Pairs
by: Xu, Heng, et al.
Published: (2024)

Defending Against Prompt Injection with DataFilter
by: Wang, Yizhu, et al.
Published: (2025)

When Safety Becomes a Vulnerability: Exploiting LLM Alignment Homogeneity for Transferable Blocking in RAG
by: Li, Junchen, et al.
Published: (2026)

Adversary-Aware DPO: Enhancing Safety Alignment in Vision Language Models via Adversarial Training
by: Weng, Fenghua, et al.
Published: (2025)

Into the Gray Zone: Domain Contexts Can Blur LLM Safety Boundaries
by: Hung, Ki Sen, et al.
Published: (2026)

Evaluating of Machine Unlearning: Robustness Verification Without Prior Modifications
by: Xu, Heng, et al.
Published: (2024)

AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning
by: Zhang, Yi, et al.
Published: (2025)

Rethinking Fraud Safety Evaluation: Multi-Round Attacks Reveal Safety-Utility Tradeoffs in Graph-Context LLM Defenders
by: Jiang, Laura, et al.
Published: (2026)

AgentGuard: Repurposing Agentic Orchestrator for Safety Evaluation of Tool Orchestration
by: Chen, Jizhou, et al.
Published: (2025)

Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable
by: Huang, Tiansheng, et al.
Published: (2025)

Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift
by: Yuan, Shuai, et al.
Published: (2025)

Adversarial Attack-Defense Co-Evolution for LLM Safety Alignment via Tree-Group Dual-Aware Search and Optimization
by: Li, Xurui, et al.
Published: (2025)

Mitigating the Safety-utility Trade-off in LLM Alignment via Adaptive Safe Context Learning
by: Wang, Yanbo, et al.
Published: (2026)

VisuoAlign: Safety Alignment of LVLMs with Multimodal Tree Search
by: Li, MingSheng, et al.
Published: (2025)

Behavioral Authentication for Security and Safety
by: Wang, Cheng, et al.
Published: (2023)

GradSafe: Detecting Jailbreak Prompts for LLMs via Safety-Critical Gradient Analysis
by: Xie, Yueqi, et al.
Published: (2024)

Benign Fine-Tuning Breaks Safety Alignment in Audio LLMs
by: Roh, Jaechul, et al.
Published: (2026)

Demystifying and Detecting Agentic Workflow Injection Vulnerabilities in GitHub Actions
by: Wang, Shenao, et al.
Published: (2026)

On Borrowed Time -- Preventing Static Side-Channel Analysis
by: Dumitru, Robert, et al.
Published: (2023)

Defending Against Neural Network Model Inversion Attacks via Data Poisoning
by: Zhou, Shuai, et al.
Published: (2024)

AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models
by: Zhang, Jinchuan, et al.
Published: (2025)

Don't Forget Too Much: Towards Machine Unlearning on Feature Level
by: Xu, Heng, et al.
Published: (2024)

What Matters For Safety Alignment?
by: Li, Xing, et al.
Published: (2026)

Targeting Alignment: Extracting Safety Classifiers of Aligned LLMs
by: Ferrand, Jean-Charles Noirot, et al.
Published: (2025)

Measuring Safety Alignment Effects in Autonomous Security Agents
by: David, Isaac, et al.
Published: (2026)