Cover Image

Understanding LLM Behavior When Encountering User-Supplied Harmful Content in Harmless Tasks

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chu, Junjie, Qu, Yiting, Leng, Ye, Backes, Michael, Shen, Yun, Zannettou, Savvas, Zhang, Yang
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns
by: Shen, Xinyue, et al.
Published: (2025)

Hate in Plain Sight: On the Risks of Moderating AI-Generated Hateful Illusions
by: Qu, Yiting, et al.
Published: (2025)

UnsafeBench: Benchmarking Image Safety Classifiers on Real-World and AI-Generated Images
by: Qu, Yiting, et al.
Published: (2024)

Breaking Agents: Compromising Autonomous LLM Agents Through Malfunction Amplification
by: Zhang, Boyang, et al.
Published: (2024)

Benchmark of Benchmarks: Unpacking Influence and Code Repository Quality in LLM Safety Benchmarks
by: Chu, Junjie, et al.
Published: (2026)

When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm
by: Leng, Ye, et al.
Published: (2026)

Prompt Stealing Attacks Against Text-to-Image Generation Models
by: Shen, Xinyue, et al.
Published: (2023)

Bridging the Gap in Vision Language Models in Identifying Unsafe Concepts Across Modalities
by: Qu, Yiting, et al.
Published: (2025)

JADES: A Universal Framework for Jailbreak Assessment via Decompositional Scoring
by: Chu, Junjie, et al.
Published: (2025)

When GPT Spills the Tea: Comprehensive Assessment of Knowledge File Leakage in GPTs
by: Shen, Xinyue, et al.
Published: (2025)

Understanding Data Importance in Machine Learning Attacks: Does Valuable Data Pose Greater Harm?
by: Wen, Rui, et al.
Published: (2024)

HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?
by: Jiang, Yukun, et al.
Published: (2026)

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

Watermarking LLM-Generated Datasets in Downstream Tasks
by: Liu, Yugeng, et al.
Published: (2025)

Robustness Over Time: Understanding Adversarial Examples' Effectiveness on Longitudinal Versions of Large Language Models
by: Liu, Yugeng, et al.
Published: (2023)

Harmless Yet Harmful: Neutral Prompting Attacks for Stealthy Hallucination Steering in Agent Skills
by: Hsu, Chia-Yi, et al.
Published: (2026)

The Trojan Knowledge: Bypassing Commercial LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search
by: Wei, Rongzhe, et al.
Published: (2025)

On the Proactive Generation of Unsafe Images From Text-To-Image Models Using Benign Prompts
by: Wu, Yixin, et al.
Published: (2023)

From Evidence to Verdict: An Agent-Based Forensic Framework for AI-Generated Image Detection
by: Liang, Mengfei, et al.
Published: (2025)

HarmRLVR: Weaponizing Verifiable Rewards for Harmful LLM Alignment
by: Liu, Yuexiao, et al.
Published: (2025)

Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency
by: Jiang, Yukun, et al.
Published: (2025)

Image-Perfect Imperfections: Safety, Bias, and Authenticity in the Shadow of Text-To-Image Model Evolution
by: Wu, Yixin, et al.
Published: (2024)

Harmless Backdoor-based Client-side Watermarking in Federated Learning
by: Luo, Kaijing, et al.
Published: (2024)

From Compression to Accountability: Harmless Copyright Protection for Dataset Distillation
by: Liang, Yan, et al.
Published: (2026)

"Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models
by: Shen, Xinyue, et al.
Published: (2023)

Synthetic Artifact Auditing: Tracing LLM-Generated Synthetic Data Usage in Downstream Applications
by: Wu, Yixin, et al.
Published: (2025)

Reconstruct Your Previous Conversations! Comprehensively Investigating Privacy Leakage Risks in Conversations with GPT Models
by: Chu, Junjie, et al.
Published: (2024)

ICLGuard: Controlling In-Context Learning Behavior for Applicability Authorization
by: Si, Wai Man, et al.
Published: (2024)

Supporting Human Raters with the Detection of Harmful Content using Large Language Models
by: Thomas, Kurt, et al.
Published: (2024)

Auditing the Compliance and Enforcement of Twitter's Advertising Policy
by: Vekaria, Yash, et al.
Published: (2023)

JailbreakRadar: Comprehensive Assessment of Jailbreak Attacks Against LLMs
by: Chu, Junjie, et al.
Published: (2024)

Pop Quiz Attack: Black-box Membership Inference Attacks Against Large Language Models
by: Chen, Zeyuan, et al.
Published: (2026)

Generated Data with Fake Privacy: Hidden Dangers of Fine-tuning Large Language Models on Generated Data
by: Akkus, Atilla, et al.
Published: (2024)

Efficient Jailbreaking of Large Models by Freeze Training: Lower Layers Exhibit Greater Sensitivity to Harmful Content
by: Shen, Hongyuan, et al.
Published: (2025)

PREE: Towards Harmless and Adaptive Fingerprint Editing in Large Language Models via Knowledge Prefix Enhancement
by: Yue, Xubin, et al.
Published: (2025)

The Challenge of Identifying the Origin of Black-Box Large Language Models
by: Yang, Ziqing, et al.
Published: (2025)

When Youth Enter the Algorithmic Wild: Discovering and Understanding Potentially Harmful Teen Videos on Douyin and Kwai
by: Zhou, Shaoxuan, et al.
Published: (2026)

Harnessing Task Overload for Scalable Jailbreak Attacks on Large Language Models
by: Dong, Yiting, et al.
Published: (2024)

Operational Runtime Behavior Mining for Open-Source Supply Chain Security
by: Tan, Zhuoran, et al.
Published: (2026)

FFT: Towards Harmlessness Evaluation and Analysis for LLMs with Factuality, Fairness, Toxicity
by: Cui, Shiyao, et al.
Published: (2023)