Surprise Alone Learns Lies: Evidence-Gated Continual Learning for LLMs from Streaming Claims

three models) but not from legitimate updates of the same kind of fact (AUC 0.53). It saves 43% of judge calls, but an attacker with access to the model

Surviving the Unseen: Predictive Defense for Novel Multi-Turn Multimodal Attacks

expansion of Multimodal Large Language Models (MLLMs) and their integration into autonomous agentic workflows has introduced a non-stationary attack surface. Empirical observations indicate that adversaries employ progressive, cross-modal

Critical-CoT: A Robust Defense Framework against Reasoning-Level Backdoor Attacks in Large Language Models

Large Language Models (LLMs), despite their impressive capabilities across domains, have been shown to be vulnerable to backdoor attacks. Prior backdoor strategies predominantly operate at the token level, where

PentestChain: A Cost-Aware, MCP-Orchestrated Framework for Automated Penetration Testing with Free-Tier LLMs

local Ollama model (qwen2.5-7b) first, then free-tier OpenRouter and Cerebras, with a rule-based fallback that always produces output-and exposes the full pipeline through a Model Context Protocol

What Guides the Agent? Adjudicating Unauthorized Behavior via Localizing Behavior-Guiding Instructions

input/output level, remains insufficient for detecting such dynamic inducements that arise during model reasoning. We propose Attnlocate, a runtime framework for fine-grained localization of context spans that genuinely influence

CrACK: Adversarial Attacks on Cross-Model Consistency in Collaborative Vision Foundation Models

Existing adversarial attacks target a single model and overlook this premise, leaving the inter-model interface entirely unguarded. We propose CrACK (Cross-model Adversarial Consistency attack), an inference-time attack

Catastrophic Learning: A New Attack Vector on Continual Learning Networks

threat model and propose six attack strategies: Label-Exchange, Tensor-Exchange, Attraction-Coincident, Attraction-Preceding, Repulsion-Coincident, and Repulsion-Preceding. The Attraction variants minimize the loss between the poisoned

Beyond Native Success: Auditing Deployment-Interface Exposure of CLIP Backdoors

Contrastive Language-Image Pre-training models are widely reused across downstream interfaces, including feature extraction, retrieval, reranking, and selection. Existing CLIP backdoor, however, usually validate attacks on a small attack

Density-aware Sample-specific Attack

derive principled criteria characterizing optimal sample-specific trigger construction under a Bayes-optimal model of the victim's training. Our analysis reveals that both attack success and clean-accuracy preservation

Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs

contexts can dramatically shift behavior outside those contexts. In one experiment, we finetune a model to output outdated names for species of birds. This causes it to behave

vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied features

DSBA: Dynamic Stealthy Backdoor Attack with Collaborative Optimization in Self-Supervised Learning

generalization capabilities, and its potential for privacy preservation. However, recent research reveals that SSL models are also vulnerable to backdoor attacks. Existing backdoor attack methods in the SSL context commonly

Color Matters: Trigger Color Affects Success in Federated Backdoor Attacks

Federated learning is vulnerable to backdoor attacks in which malicious clients inject poisoned updates while preserving benign-task performance. In this paper, we study a semantics-driven backdoor mechanism

Detecting Trojaned DNNs via Spectral Regression Analysis

approach that analyzes how a model's internal representations change during fine-tuning. Rather than attempting to reconstruct trigger conditions, MIST characterizes benign model evolution using pre-activation spectra

Agentic AI as a Cybersecurity Attack Surface: Threats, Exploits, and Defenses in Runtime Supply Chains

Agentic systems built on large language models (LLMs) extend beyond text generation to autonomously retrieve information and invoke tools. This runtime execution model shifts the attack surface from build-time

Security of World-Model-Based Embodied AI: A Lifecycle of Threats, Defenses, and Evaluation

World models give embodied AI a predictive core: they compress observations into states, simulate action-conditioned futures, and enable planning beyond reactive control. This predictive layer, however, opens

Deep Learning Latency Attacks and Defenses: A Cross-Domain Survey of Availability Threats

verbose-image, and reasoning denial-of-service attacks on LLMs, VLMs, mixture-of-experts models, and tool-using agents). We organize attacks by exploited computational bottleneck rather than formulation, separating

Autonomous OSS Threat Detection via Taxonomy-Aligned LLMs

attacks including typosquatting, dependency confusion, Trojan Source obfuscation, malicious build injection, and CI/CD pipeline poisoning. Existing detection approaches rely on signature-based tools and rule-based systems that struggle

ProjLens: Unveiling the Role of Projectors in Multimodal Model Safety

Multimodal Large Language Models (MLLMs) have achieved remarkable success in cross-modal understanding and generation, yet their deployment is threatened by critical safety vulnerabilities. While prior works have demonstrated

LoRA as Oracle

Existing defenses for backdoor detection and membership inference typically require access to clean reference models, extensive retraining, or strong assumptions about the attack mechanism. In this work, we introduce