Imagen de Portada

LLM-Safety Evaluations Lack Robustness

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Beyer, Tim, Xhonneux, Sophie, Geisler, Simon, Gidel, Gauthier, Schwinn, Leo, Günnemann, Stephan
Formato:	Preprint
Publicado:	2025
Materias:	Cryptography and Security Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Fast Proxies for LLM Robustness Evaluation
por: Beyer, Tim, et al.
Publicado: (2025)

Efficient Adversarial Training in LLMs with Continuous Attacks
por: Xhonneux, Sophie, et al.
Publicado: (2024)

A Generative Approach to LLM Harmfulness Mitigation with Red Flag Tokens
por: Dobre, David, et al.
Publicado: (2025)

A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness
por: Schwinn, Leo, et al.
Publicado: (2026)

Revisiting the Robust Alignment of Circuit Breakers
por: Schwinn, Leo, et al.
Publicado: (2024)

Soft Prompt Threats: Attacking Safety Alignment and Unlearning in Open-Source LLMs through the Embedding Space
por: Schwinn, Leo, et al.
Publicado: (2024)

In-Context Learning Can Re-learn Forbidden Tasks
por: Xhonneux, Sophie, et al.
Publicado: (2024)

Model Collapse Is Not a Bug but a Feature in Machine Unlearning for LLMs
por: Scholten, Yan, et al.
Publicado: (2025)

Closing the Distribution Gap in Adversarial Training for LLMs
por: Hu, Chengzhi, et al.
Publicado: (2026)

SAGE: A Generic Framework for LLM Safety Evaluation
por: Jindal, Madhur, et al.
Publicado: (2025)

AdversariaLLM: A Unified and Modular Toolbox for LLM Robustness Research
por: Beyer, Tim, et al.
Publicado: (2025)

Adaptive and Robust Cost-Aware Proof of Quality for Decentralized LLM Inference Networks
por: Tian, Arther, et al.
Publicado: (2026)

ShadowLogic: Backdoors in Any Whitebox LLM
por: Schulz, Kasimir, et al.
Publicado: (2025)

Evaluating Answer Leakage Robustness of LLM Tutors against Adversarial Student Attacks
por: Zhao, Jin, et al.
Publicado: (2026)

aiXamine: Simplified LLM Safety and Security
por: Deniz, Fatih, et al.
Publicado: (2025)

FreakOut-LLM: The Effect of Emotional Stimuli on Safety Alignment
por: Kuznetsov, Daniel, et al.
Publicado: (2026)

PRISM: Robust VLM Alignment with Principled Reasoning for Integrated Safety in Multimodality
por: Li, Nanxi, et al.
Publicado: (2025)

Safety Layers in Aligned Large Language Models: The Key to LLM Security
por: Li, Shen, et al.
Publicado: (2024)

QGuard:Question-based Zero-shot Guard for Multi-modal LLM Safety
por: Lee, Taegyeong, et al.
Publicado: (2025)

Quantifying LLM Safety Degradation Under Repeated Attacks Using Survival Analysis
por: Topol, Zvi
Publicado: (2026)

TamperBench: Systematically Stress-Testing LLM Safety Under Fine-Tuning and Tampering
por: Hossain, Saad, et al.
Publicado: (2026)

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

SFCoT: Safer Chain-of-Thought via Active Safety Evaluation and Calibration
por: Pan, Yu, et al.
Publicado: (2026)

TED-LaST: Towards Robust Backdoor Defense Against Adaptive Attacks
por: Mo, Xiaoxing, et al.
Publicado: (2025)

FPEdit: Robust LLM Fingerprinting through Localized Parameter Editing
por: Wang, Shida, et al.
Publicado: (2025)

AgentTrust: Runtime Safety Evaluation and Interception for AI Agent Tool Use
por: Yang, Chenglin
Publicado: (2026)

Character-Level Perturbations Disrupt LLM Watermarks
por: Zhang, Zhaoxi, et al.
Publicado: (2025)

LLM-based event log analysis techniques: A survey
por: Akhtar, Siraaj, et al.
Publicado: (2025)

NCCR: to Evaluate the Robustness of Neural Networks and Adversarial Examples
por: Pu, Shi, et al.
Publicado: (2025)

USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models
por: Zheng, Baolin, et al.
Publicado: (2025)

Efficient LLM Safety Evaluation through Multi-Agent Debate
por: Lin, Dachuan, et al.
Publicado: (2025)

SafeGenes: Evaluating the Adversarial Robustness of Genomic Foundation Models
por: Zhan, Huixin, et al.
Publicado: (2025)

Evaluating and Mitigating LLM-as-a-judge Bias in Communication Systems
por: Gao, Jiaxin, et al.
Publicado: (2025)

Seclens: Role-specific Evaluation of LLM's for security vulnerablity detection
por: Halder, Subho, et al.
Publicado: (2026)

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

CyberLLMInstruct: A Pseudo-malicious Dataset Revealing Safety-performance Trade-offs in Cyber Security LLM Fine-tuning
por: ElZemity, Adel, et al.
Publicado: (2025)

COGNITION: From Evaluation to Defense against Multimodal LLM CAPTCHA Solvers
por: Wang, Junyu, et al.
Publicado: (2025)

Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities
por: Che, Zora, et al.
Publicado: (2025)

Taxonomy, Evaluation and Exploitation of IPI-Centric LLM Agent Defense Frameworks
por: Ji, Zimo, et al.
Publicado: (2025)

Supporting Artifact Evaluation with LLMs: A Study with Published Security Research Papers
por: Heye, David, et al.
Publicado: (2026)