Cover Image

COGNITION: From Evaluation to Defense against Multimodal LLM CAPTCHA Solvers

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wang, Junyu, Zhu, Changjia, Zhou, Yuanbo, Li, Lingyao, He, Xu, Wei, Mingkui, Xiong, Junjie
Format:	Preprint
Published:	2025
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

Invisible Prompts, Visible Threats: Malicious Font Injection in External Resources for Large Language Models
by: Xiong, Junjie, et al.
Published: (2025)

Spatial CAPTCHA: Generatively Benchmarking Spatial Reasoning for Human-Machine Differentiation
by: Kharlamova, Arina, et al.
Published: (2025)

Reverse CAPTCHA: Evaluating LLM Susceptibility to Invisible Unicode Instruction Injection
by: Graves, Marcus
Published: (2026)

When Your Reviewer is an LLM: Biases, Divergence, and Prompt Injection Risks in Peer Review
by: Zhu, Changjia, et al.
Published: (2025)

Oedipus: LLM-enchanced Reasoning CAPTCHA Solver
by: Deng, Gelei, et al.
Published: (2024)

aCAPTCHA: Verifying That an Entity Is a Capable Agent via Asymmetric Hardness
by: Xu, Zuyao, et al.
Published: (2026)

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

Secure Tug-of-War (SecTOW): Iterative Defense-Attack Training with Reinforcement Learning for Multimodal Model Security
by: Dai, Muzhi, et al.
Published: (2025)

UNSEEN: A Cross-Stack LLM Unlearning Defense against AR-LLM Social Engineering Attacks
by: Yu, Tianlong, et al.
Published: (2026)

A Hybrid CAPTCHA Combining Generative AI with Keystroke Dynamics for Enhanced Bot Detection
by: Nia, Ayda Aghaei
Published: (2025)

SoK: The Last Line of Defense: On Backdoor Defense Evaluation
by: Abad, Gorka, et al.
Published: (2025)

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

Constitutional Classifiers++: Efficient Production-Grade Defenses against Universal Jailbreaks
by: Cunningham, Hoagy, et al.
Published: (2026)

Ensemble Privacy Defense for Knowledge-Intensive LLMs against Membership Inference Attacks
by: Fu, Haowei, et al.
Published: (2025)

A Survey on Backdoor Threats in Large Language Models (LLMs): Attacks, Defenses, and Evaluations
by: Zhou, Yihe, et al.
Published: (2025)

SDD: Self-Degraded Defense against Malicious Fine-tuning
by: Chen, Zixuan, et al.
Published: (2025)

BackdoorMBTI: A Backdoor Learning Multimodal Benchmark Tool Kit for Backdoor Defense Evaluation
by: Yu, Haiyang, et al.
Published: (2024)

FraudShield: Knowledge Graph Empowered Defense for LLMs against Fraud Attacks
by: Xu, Naen, et al.
Published: (2026)

MARS: A Malignity-Aware Backdoor Defense in Federated Learning
by: Wan, Wei, et al.
Published: (2025)

I Don't Know You, But I Can Catch You: Real-Time Defense against Diverse Adversarial Patches for Object Detectors
by: Lin, Zijin, et al.
Published: (2024)

Multimodal Prompt Injection Attacks: Risks and Defenses for Modern LLMs
by: Yeo, Andrew, et al.
Published: (2025)

Model Inversion Attack against Federated Unlearning
by: Zhou, Lei, et al.
Published: (2025)

Exploring Backdoor Attack and Defense for LLM-empowered Recommendations
by: Ning, Liangbo, et al.
Published: (2025)

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

Automated Framework to Evaluate and Harden LLM System Instructions against Encoding Attacks
by: Sahu, Anubhab, et al.
Published: (2026)

Surviving the Unseen: Predictive Defense for Novel Multi-Turn Multimodal Attacks
by: You, Doohee
Published: (2026)

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory
by: Wei, Qianshan, et al.
Published: (2025)

BELT: Old-School Backdoor Attacks can Evade the State-of-the-Art Defense with Backdoor Exclusivity Lifting
by: Qiu, Huming, et al.
Published: (2023)

BounTCHA: A CAPTCHA Utilizing Boundary Identification in Guided Generative AI-extended Videos
by: Lin, Lehao, et al.
Published: (2025)

Evaluation of Prompt Injection Defenses in Large Language Models
by: Deep, Priyal, et al.
Published: (2026)

DistillGuard: Evaluating Defenses Against LLM Knowledge Distillation
by: Jiang, Bo
Published: (2026)

Operationalizing CaMeL: Strengthening LLM Defenses for Enterprise Deployment
by: Tallam, Krti, et al.
Published: (2025)

The Best Defense is a Good Offense: Countering LLM-Powered Cyberattacks
by: Ayzenshteyn, Daniel, et al.
Published: (2024)

Class-Conditional Neural Polarizer: A Lightweight and Effective Backdoor Defense by Purifying Poisoned Features
by: Zhu, Mingli, 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)

Recent Advances in Attack and Defense Approaches of Large Language Models
by: Cui, Jing, et al.
Published: (2024)

KG-DF: A Black-box Defense Framework against Jailbreak Attacks Based on Knowledge Graphs
by: Liu, Shuyuan, et al.
Published: (2025)

From Allies to Adversaries: Manipulating LLM Tool-Calling through Adversarial Injection
by: Wang, Haowei, et al.
Published: (2024)

SHIELD: An Auto-Healing Agentic Defense Framework for LLM Resource Exhaustion Attacks
by: Sivaroopan, Nirhoshan, et al.
Published: (2026)

DRIFT: Dynamic Rule-Based Defense with Injection Isolation for Securing LLM Agents
by: Li, Hao, et al.
Published: (2025)