SelfGrader: LLM Jailbreak Detection via Anchored Token-Level Logits
Explorar
Noticias de IA
1359 elementos — filtrados, clasificados y sin duplicados
Fed up with vibe coders, dev sneaks data-nuking prompt injection into their code
SPARD: Defending Harmful Fine-Tuning Attack via Safety Projection with Relevance-Diversit…
Disentangling Adversarial Prompts: A Semantic-Graph Defense for Robust LLM Security
SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents
Backdoor Attacks on Fault Detection and Localization in Cyber-Physical Systems
Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems
Semantic-level Backdoor Attack against Text-to-Image Diffusion Models
HARP: Measuring Harm Amplification in Multi-Agent LLM Systems
Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models
The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes
Blind PRNG Hijacking: An Undetectable Integrity-Preserving Attack Against LLM Watermarking
Symmetry Defeats Auditing
Defending LLM-based Multi-Agent Systems Against Cooperative Attacks with Sentence-Level R…
Technical Report: Exploring the Emerging Threats of the Agent Skill Ecosystem
Grimlock: Guarding High-Agency Systems with eBPF and Attested Channels
MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Conte…
Securing Retrieval-Augmented Generation: A Taxonomy of Attacks, Defenses, and Future Dire…
Plant, Persist, Trigger: Sleeper Attack on Large Language Model Agents
A Wolf in Sheep's Clothing: Targeted Routing Hijacking in Federated RAG
Defending LLM-based Multi-Agent Systems Against Cooperative Attacks with Sentence-Level R…
SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adv…
Indian Government, Tech Firms Running Tests for Mythos Threat
Cordon-MAS: Defending RAG against Knowledge Poisoning via Information-Flow Control
Cryptographic Registry Provenance: Structural Defense Against Dependency Confusion in AI …
GraphIP-Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It?
Securing Multi-Agent Systems Against Corruptions via Node Contribution Backpropagation
MemMorph: Tool Hijacking in LLM Agents via Memory Poisoning
Erased but Exploitable: Black-box Embedding-Aware Prompting Against Unlearned Text-to-Ima…
Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception