Eyes-on-Me: Scalable RAG Poisoning through Transferable Attention-Steering Attractors

data poisoning and show that modular, reusable components pose a practical threat to modern AI systems. They also reveal a strong link between attention concentration and model outputs, informing interpretability

CODE: A Contradiction-Based Deliberation Extension Framework for Overthinking Attacks on Retrieval-Augmented Generation

multi-step self-verification. However, recent studies have shown that reasoning models suffer from overthinking attacks, where models are tricked to generate unnecessarily high number of reasoning tokens. In this

Assessing Attack Surfaces in Generative Search Engines through Publisher Attributes: A Case Study in Political Domains

language models (LLMs). They play a crucial role in how users access information on the web. Because anyone can publish content on the web, GSEs are vulnerable to poisoning attacks

Log-To-Leak: Prompt Injection Attacks on Tool-Using LLM Agents via the Model Context Protocol

tools through the Model Context Protocol (MCP), but the natural-language metadata used to describe those tools creates a prompt-injection surface. Within this established tool-poisoning surface, we study

Security in the Fine-Tuning Lifecycle of Large Language Models: Threats, Defenses,Evaluation, and Future Directions

Language Models (LLMs) to downstream tasks, but its reliance on training data, parameter updates, and reusable components opens entry points for attackers. Threats have evolved from data poisoning and weight

Breadcrumbing Search Agents

single poisoned page, a mediated search interface can repeatedly steer how the agent gathers evidence and forms its final answer. Under a constrained tool-intermediary threat model, appending only

Adaptive Defense Orchestration for RAG: A Sentinel-Strategist Architecture against Multi-Vector Attacks

poisoning, the strongest ADO variants reduce attack success to near zero while restoring contextual recall to more than 75% of the undefended baseline, although robustness remains sensitive to model choice

Evaluating Adversarial Attacks on Federated Learning for Temperature Forecasting

high-resolution spatiotemporal forecasts that can surpass traditional numerical models, while FL allows institutions in different locations to collaboratively train models without sharing raw data, addressing efficiency and security concerns

Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

reliably backdoored through training data poisoning while preserving baseline task performance. On a Qwen 2.5 1.5B prompt-injection classifier, a small fraction of poisoned examples drives a clean-accuracy

SecureSplit: Mitigating Backdoor Attacks in Split Learning

trained model. To address this vulnerability, we introduce SecureSplit, a defense mechanism tailored to SL. SecureSplit applies a dimensionality transformation strategy to accentuate subtle differences between benign and poisoned embeddings

FedGuard-DC: Privacy-Preserving Federated Load Forecasting and Cyber-Attack Detection for Data-Center Loads in Transmission Systems

while only model updates are shared with the global controller. Optional differential privacy and robust trimmed-mean aggregation are included to evaluate privacy-utility behavior and poisoned-client resilience

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

perplexity. Cross-model evaluation across four target LLMs shows nontrivial effectiveness under a fixed trigger generator, and transfer tests against unseen retrievers, including ColBERT and commercial embedding models, yield

Multi-Targeted Graph Backdoor Attack

based attack. Our analysis on four GNN models confirms the generalization capability of our attack which is effective regardless of the GNN model architectures and training parameters settings. We further

RIFLE: Robust Distillation-based FL for Deep Model Deployment on Resource-Constrained IoT Networks

TinyML models, collaboratively train global models by sharing gradients with a central server while preserving data privacy. However, as data heterogeneity and task complexity increase, TinyML models often become insufficient

Open WebUI: Redis Cache Keys tool_servers and terminal_servers

A Bag of Tricks for Defending Deep Learning-based Face Recognition Pipelines against Backdoor Attacks

crucial to system providers. This technical report targets the threat model most relevant to industry practitioners: although model training is performed in-house, training data is sourced from possibly-compromised

Rescuing the Unpoisoned: Efficient Defense against Knowledge Corruption Attacks on RAG Systems

poisoning) attacks in practical RAG deployments. RAGDefender operates during the post-retrieval phase, leveraging lightweight machine learning techniques to detect and filter out adversarial content without requiring additional model training

Robustness of Anomaly Detection Models for Industrial Control Systems under Training-Time Data Contamination

poisoning methods. Contamination budgets from 1% to 10% are evaluated using clean validation and test sets under a unified offline protocol. The results show that robustness is strongly model-dependent

Stealing Reasoning Traces from Proprietary LLM APIs

jailbreaking the more capable model directly. This vulnerability enables four distinct attack vectors. First, it circumvents anti-distillation mechanisms, allowing adversaries to extract a proprietary model's reasoning

Safety, Security, and Cognitive Risks in Neuro-Symbolic AI

threat model extending MITRE ATLAS with 11 NeSy-specific tactic extensions and a five-profile attacker taxonomy; (3) a symbolic-layer threat catalogue covering knowledge graph (KG) poisoning, ontology-merging