Imagen de Portada

SafeDecoding: Defending against Jailbreak Attacks via Safety-Aware Decoding

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Xu, Zhangchen, Jiang, Fengqing, Niu, Luyao, Jia, Jinyuan, Lin, Bill Yuchen, Poovendran, Radha
Formato:	Preprint
Publicado:	2024
Materias:	Cryptography and Security Artificial Intelligence Computation and Language
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

ChatBug: A Common Vulnerability of Aligned LLMs Induced by Chat Templates
por: Jiang, Fengqing, et al.
Publicado: (2024)

ACE: A Model Poisoning Attack on Contribution Evaluation Methods in Federated Learning
por: Xu, Zhangchen, et al.
Publicado: (2024)

Brave: Byzantine-Resilient and Privacy-Preserving Peer-to-Peer Federated Learning
por: Xu, Zhangchen, et al.
Publicado: (2024)

CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models
por: Li, Yuetai, et al.
Publicado: (2024)

SoSBench: Benchmarking Safety Alignment on Six Scientific Domains
por: Jiang, Fengqing, et al.
Publicado: (2025)

BadScientist: Can a Research Agent Write Convincing but Unsound Papers that Fool LLM Reviewers?
por: Jiang, Fengqing, et al.
Publicado: (2025)

Stronger Models are NOT Stronger Teachers for Instruction Tuning
por: Xu, Zhangchen, et al.
Publicado: (2024)

ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs
por: Jiang, Fengqing, et al.
Publicado: (2024)

SafeAligner: Safety Alignment against Jailbreak Attacks via Response Disparity Guidance
por: Huang, Caishuang, et al.
Publicado: (2024)

SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities
por: Jiang, Fengqing, et al.
Publicado: (2025)

BadChain: Backdoor Chain-of-Thought Prompting for Large Language Models
por: Xiang, Zhen, et al.
Publicado: (2024)

Game of Trojans: Adaptive Adversaries Against Output-based Trojaned-Model Detectors
por: Sahabandu, Dinuka, et al.
Publicado: (2024)

RobustKV: Defending Large Language Models against Jailbreak Attacks via KV Eviction
por: Jiang, Tanqiu, et al.
Publicado: (2024)

JBShield: Defending Large Language Models from Jailbreak Attacks through Activated Concept Analysis and Manipulation
por: Zhang, Shenyi, et al.
Publicado: (2025)

AISA: Awakening Intrinsic Safety Awareness in Large Language Models against Jailbreak Attacks
por: Song, Weiming, et al.
Publicado: (2026)

HSF: Defending against Jailbreak Attacks with Hidden State Filtering
por: Qian, Cheng, et al.
Publicado: (2024)

Defending against Backdoor Attack on Deep Neural Networks
por: Cheng, Hao, et al.
Publicado: (2020)

Adversarial Tuning: Defending Against Jailbreak Attacks for LLMs
por: Liu, Fan, et al.
Publicado: (2024)

SelfDefend: LLMs Can Defend Themselves against Jailbreaking in a Practical Manner
por: Wang, Xunguang, et al.
Publicado: (2024)

TASO: Jailbreak LLMs via Alternative Template and Suffix Optimization
por: Wang, Yanting, et al.
Publicado: (2025)

Defending against Data Poisoning Attacks in Federated Learning via User Elimination
por: Galanis, Nick
Publicado: (2024)

PandaGuard: Systematic Evaluation of LLM Safety against Jailbreaking Attacks
por: Shen, Guobin, et al.
Publicado: (2025)

Provably Robust Explainable Graph Neural Networks against Graph Perturbation Attacks
por: Li, Jiate, et al.
Publicado: (2025)

Robustness via Referencing: Defending against Prompt Injection Attacks by Referencing the Executed Instruction
por: Chen, Yulin, et al.
Publicado: (2025)

TinyV: Reducing False Negatives in Verification Improves RL for LLM Reasoning
por: Xu, Zhangchen, et al.
Publicado: (2025)

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
por: Xu, Zhangchen, et al.
Publicado: (2024)

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

Defending Jailbreak Prompts via In-Context Adversarial Game
por: Zhou, Yujun, et al.
Publicado: (2024)

SafeDream: Safety World Model for Proactive Early Jailbreak Detection
por: Yan, Bo, et al.
Publicado: (2026)

Invariant Aggregator for Defending against Federated Backdoor Attacks
por: Wang, Xiaoyang, et al.
Publicado: (2022)

Concept-Aware Privacy Mechanisms for Defending Embedding Inversion Attacks
por: Tsai, Yu-Che, et al.
Publicado: (2026)

A Critical Evaluation of Defenses against Prompt Injection Attacks
por: Jia, Yuqi, et al.
Publicado: (2025)

MetaDefense: Defending Finetuning-based Jailbreak Attack Before and During Generation
por: Jiang, Weisen, et al.
Publicado: (2025)

CANTXSec: A Deterministic Intrusion Detection and Prevention System for CAN Bus Monitoring ECU Activations
por: Donadel, Denis, et al.
Publicado: (2025)

RePD: Defending Jailbreak Attack through a Retrieval-based Prompt Decomposition Process
por: Wang, Peiran, et al.
Publicado: (2024)

Double-Dip: Thwarting Label-Only Membership Inference Attacks with Transfer Learning and Randomization
por: Rajabi, Arezoo, et al.
Publicado: (2024)

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

AutoJailbreak: Exploring Jailbreak Attacks and Defenses through a Dependency Lens
por: Lu, Lin, et al.
Publicado: (2024)

Defensive Prompt Patch: A Robust and Interpretable Defense of LLMs against Jailbreak Attacks
por: Xiong, Chen, et al.
Publicado: (2024)

RL-JACK: Reinforcement Learning-powered Black-box Jailbreaking Attack against LLMs
por: Chen, Xuan, et al.
Publicado: (2024)