Copertina

AdvBDGen: Adversarially Fortified Prompt-Specific Fuzzy Backdoor Generator Against LLM Alignment

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Pathmanathan, Pankayaraj, Sehwag, Udari Madhushani, Panaitescu-Liess, Michael-Andrei, Huang, Furong
Natura:	Preprint
Pubblicazione:	2024
Soggetti:	Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

RAGPart & RAGMask: Retrieval-Stage Defenses Against Corpus Poisoning in Retrieval-Augmented Generation
di: Pathmanathan, Pankayaraj, et al.
Pubblicazione: (2025)

Deliberative Alignment is Deep, but Uncertainty Remains: Inference time safety improvement in reasoning via attribution of unsafe behavior to base model
di: Pathmanathan, Pankayaraj, et al.
Pubblicazione: (2026)

PoisonedParrot: Subtle Data Poisoning Attacks to Elicit Copyright-Infringing Content from Large Language Models
di: Panaitescu-Liess, Michael-Andrei, et al.
Pubblicazione: (2025)

Can Watermarking Large Language Models Prevent Copyrighted Text Generation and Hide Training Data?
di: Panaitescu-Liess, Michael-Andrei, et al.
Pubblicazione: (2024)

Reward Models Can Improve Themselves: Reward-Guided Adversarial Failure Mode Discovery for Robust Reward Modeling
di: Pathmanathan, Pankayaraj, et al.
Pubblicazione: (2025)

Automatic Pseudo-Harmful Prompt Generation for Evaluating False Refusals in Large Language Models
di: An, Bang, et al.
Pubblicazione: (2024)

PropensityBench: Evaluating Latent Safety Risks in Large Language Models via an Agentic Approach
di: Sehwag, Udari Madhushani, et al.
Pubblicazione: (2025)

Compositional Adversarial Training for Robust Visual Watermarking
di: Satheesh, Anirudh, et al.
Pubblicazione: (2026)

Like Oil and Water: Group Robustness Methods and Poisoning Defenses May Be at Odds
di: Panaitescu-Liess, Michael-Andrei, et al.
Pubblicazione: (2025)

Is poisoning a real threat to LLM alignment? Maybe more so than you think
di: Pathmanathan, Pankayaraj, et al.
Pubblicazione: (2024)

PerceptionCLIP: Visual Classification by Inferring and Conditioning on Contexts
di: An, Bang, et al.
Pubblicazione: (2023)

GenARM: Reward Guided Generation with Autoregressive Reward Model for Test-time Alignment
di: Xu, Yuancheng, et al.
Pubblicazione: (2024)

ASPI: Seeking Ambiguity Clarification Amplifies Prompt Injection Vulnerability in LLM Agents
di: Sehwag, Udari Madhushani, et al.
Pubblicazione: (2026)

Collab: Controlled Decoding using Mixture of Agents for LLM Alignment
di: Chakraborty, Souradip, et al.
Pubblicazione: (2025)

In-Context Learning with Topological Information for Knowledge Graph Completion
di: Sehwag, Udari Madhushani, et al.
Pubblicazione: (2024)

Using Curiosity for an Even Representation of Tasks in Continual Offline Reinforcement Learning
di: Pathmanathan, Pankayaraj, et al.
Pubblicazione: (2023)

Can LLMs be Scammed? A Baseline Measurement Study
di: Sehwag, Udari Madhushani, et al.
Pubblicazione: (2024)

LHAW: Controllable Underspecification for Long-Horizon Tasks
di: Pu, George, et al.
Pubblicazione: (2026)

Defensive Refusal Bias: How Safety Alignment Fails Cyber Defenders
di: Campbell, David, et al.
Pubblicazione: (2026)

Continual Learning of Domain Knowledge from Human Feedback in Text-to-SQL
di: Cook, Thomas, et al.
Pubblicazione: (2025)

Krait: A Backdoor Attack Against Graph Prompt Tuning
di: Song, Ying, et al.
Pubblicazione: (2024)

PromptFix: Few-shot Backdoor Removal via Adversarial Prompt Tuning
di: Zhang, Tianrong, et al.
Pubblicazione: (2024)

Rethinking Adversarial Policies: A Generalized Attack Formulation and Provable Defense in RL
di: Liu, Xiangyu, et al.
Pubblicazione: (2023)

Adversarial Feature Map Pruning for Backdoor
di: Huang, Dong, et al.
Pubblicazione: (2023)

AgentCrypt: Advancing Privacy and (Secure) Computation in AI Agent Collaboration
di: Karthikeyan, Harish, et al.
Pubblicazione: (2025)

Adapting to Evolving Adversaries with Regularized Continual Robust Training
di: Dai, Sihui, et al.
Pubblicazione: (2025)

AdvBlur: Adversarial Blur for Robust Diabetic Retinopathy Classification and Cross-Domain Generalization
di: Kanagalingam, Heethanjan, et al.
Pubblicazione: (2025)

Composite Backdoor Attacks Against Large Language Models
di: Huang, Hai, et al.
Pubblicazione: (2023)

Backdoor or Manipulation? Graph Mixture of Experts Can Defend Against Various Graph Adversarial Attacks
di: Feng, Yuyuan, et al.
Pubblicazione: (2025)

Transfer Q Star: Principled Decoding for LLM Alignment
di: Chakraborty, Souradip, et al.
Pubblicazione: (2024)

Adversarially Guided Stateful Defense Against Backdoor Attacks in Federated Deep Learning
di: Ali, Hassan, et al.
Pubblicazione: (2024)

Fortifying Fully Convolutional Generative Adversarial Networks for Image Super-Resolution Using Divergence Measures
di: Basu, Arkaprabha, et al.
Pubblicazione: (2024)

On the Robustness of Graph Reduction Against GNN Backdoor
di: Zhu, Yuxuan, et al.
Pubblicazione: (2024)

AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs
di: Paulus, Anselm, et al.
Pubblicazione: (2024)

Instruction Backdoor Attacks Against Customized LLMs
di: Zhang, Rui, et al.
Pubblicazione: (2024)

Adversarial Prompt Evaluation: Systematic Benchmarking of Guardrails Against Prompt Input Attacks on LLMs
di: Zizzo, Giulio, et al.
Pubblicazione: (2025)

Poisoning the Watchtower: Prompt Injection Attacks Against LLM-Augmented Security Operations Through Adversarial Log Content
di: Pandey, Rohan, et al.
Pubblicazione: (2026)

The Task Shield: Enforcing Task Alignment to Defend Against Indirect Prompt Injection in LLM Agents
di: Jia, Feiran, et al.
Pubblicazione: (2024)

Boosting Graph Robustness Against Backdoor Attacks: An Over-Similarity Perspective
di: Liu, Chang, et al.
Pubblicazione: (2025)

Enhance Modality Robustness in Text-Centric Multimodal Alignment with Adversarial Prompting
di: Tsai, Yun-Da, et al.
Pubblicazione: (2024)