Imagen de Portada

Enhancing Agent Safety Judgment: Controlled Benchmark Rewriting and Analogical Reasoning for Deceptive Out-of-Distribution Scenarios

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Zhang, Zuoyu, Zhu, Yancheng
Formato:	Preprint
Publicado:	2026
Materias:	Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

DeceptionBench: A Comprehensive Benchmark for AI Deception Behaviors in Real-world Scenarios
por: Huang, Yao, et al.
Publicado: (2025)

Enhancing Tool Calling in LLMs with the International Tool Calling Dataset
por: Zhang, Zuoyu, et al.
Publicado: (2026)

Intentional Deception as Controllable Capability in LLM Agents
por: Starace, Jason, et al.
Publicado: (2026)

LegalReasoner: Step-wised Verification-Correction for Legal Judgment Reasoning
por: Shi, Weijie, et al.
Publicado: (2025)

Deceptive Risk Minimization: Out-of-Distribution Generalization by Deceiving Distribution Shift Detectors
por: Majumdar, Anirudha
Publicado: (2025)

Elicit and Enhance: Advancing Multimodal Reasoning in Medical Scenarios
por: Huang, Zhongzhen, et al.
Publicado: (2025)

AMS-IO-Bench and AMS-IO-Agent: Benchmarking and Structured Reasoning for Analog and Mixed-Signal Integrated Circuit Input/Output Design
por: Zhang, Zhishuai, et al.
Publicado: (2025)

Towards Benchmarking and Assessing the Safety and Robustness of Autonomous Driving on Safety-critical Scenarios
por: Li, Jingzheng, et al.
Publicado: (2025)

Safety2Drive: Safety-Critical Scenario Benchmark for the Evaluation of Autonomous Driving
por: Li, Jingzheng, et al.
Publicado: (2025)

AgentDrive: An Open Benchmark Dataset for Agentic AI Reasoning with LLM-Generated Scenarios in Autonomous Systems
por: Ferrag, Mohamed Amine, et al.
Publicado: (2026)

LPS-Bench: Benchmarking Safety Awareness of Computer-Use Agents in Long-Horizon Planning under Benign and Adversarial Scenarios
por: Chen, Tianyu, et al.
Publicado: (2026)

CARV: A Diagnostic Benchmark for Compositional Analogical Reasoning in Multimodal LLMs
por: Du, Yongkang, et al.
Publicado: (2026)

AI Deception: Risks, Dynamics, and Controls
por: Chen, Boyuan, et al.
Publicado: (2025)

CausalFlip: A Benchmark for LLM Causal Judgment Beyond Semantic Matching
por: Wang, Yuzhe, et al.
Publicado: (2026)

Out-of-Distribution Detection for Safety Assurance of AI and Autonomous Systems
por: Hodge, Victoria J., et al.
Publicado: (2025)

Chinese SafetyQA: A Safety Short-form Factuality Benchmark for Large Language Models
por: Tan, Yingshui, et al.
Publicado: (2024)

MobilityBench: A Benchmark for Evaluating Route-Planning Agents in Real-World Mobility Scenarios
por: Song, Zhiheng, et al.
Publicado: (2026)

MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents
por: Chen, Ruihan, et al.
Publicado: (2025)

Exploring the Necessity of Reasoning in LLM-based Agent Scenarios
por: Zhou, Xueyang, et al.
Publicado: (2025)

FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality
por: Chen, Keyu, et al.
Publicado: (2024)

Benchmarking MLLM-based Web Understanding: Reasoning, Robustness and Safety
por: Liu, Junliang, et al.
Publicado: (2025)

CrashAgent: Crash Scenario Generation via Multi-modal Reasoning
por: Li, Miao, et al.
Publicado: (2025)

AnalogAgent: Self-Improving Analog Circuit Design Automation with LLM Agents
por: Bao, Zhixuan, et al.
Publicado: (2026)

The Agent's First Day: Benchmarking Learning, Exploration, and Scheduling in the Workplace Scenarios
por: Fu, Daocheng, et al.
Publicado: (2026)

Embedding Trajectory for Out-of-Distribution Detection in Mathematical Reasoning
por: Wang, Yiming, et al.
Publicado: (2024)

Retrieval- and Argumentation-Enhanced Multi-Agent LLMs for Judgmental Forecasting (Extended Version with Supplementary Material)
por: Gorur, Deniz, et al.
Publicado: (2025)

AgentEscapeBench: Evaluating Out-of-Domain Tool-Grounded Reasoning in LLM Agents
por: Guo, Zhengkang, et al.
Publicado: (2026)

CR4T: Rewrite-Based Guardrails for Adolescent LLM Safety
por: An, Heajun, et al.
Publicado: (2026)

DecepChain: Inducing Deceptive Reasoning in Large Language Models
por: Shen, Wei, et al.
Publicado: (2025)

Reasoning Court: Combining Reasoning, Action, and Judgment for Multi-Hop Reasoning
por: Wu, Jingtian, et al.
Publicado: (2025)

UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios
por: Luo, Haotian, et al.
Publicado: (2025)

SafeMind: Benchmarking and Mitigating Safety Risks in Embodied LLM Agents
por: Chen, Ruolin, et al.
Publicado: (2025)

A Distribution Semantics for Probabilistic Term Rewriting
por: Vidal, Germán
Publicado: (2024)

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
por: Hubinger, Evan, et al.
Publicado: (2024)

OpenDeception: Learning Deception and Trust in Human-AI Interaction via Multi-Agent Simulation
por: Wu, Yichen, et al.
Publicado: (2025)

LABSHIELD: A Multimodal Benchmark for Safety-Critical Reasoning and Planning in Scientific Laboratories
por: Sun, Qianpu, et al.
Publicado: (2026)

I-RAVEN-X: Benchmarking Generalization and Robustness of Analogical and Mathematical Reasoning in Large Language and Reasoning Models
por: Camposampiero, Giacomo, et al.
Publicado: (2025)

OptAgent: Optimizing Query Rewriting for E-commerce via Multi-Agent Simulation
por: Handa, Divij, et al.
Publicado: (2025)

Don't Click That: Teaching Web Agents to Resist Deceptive Interfaces
por: Zhang, Yilin, et al.
Publicado: (2026)

Cognitive Inception: Agentic Reasoning against Visual Deceptions by Injecting Skepticism
por: Zhao, Yinjie, et al.
Publicado: (2025)