Category
1,407 papers
Overall, the results show that SNN design choice, evaluation protocol, spike-timing representation, and computational activity should be examined separately when applying SNNs to static network-flow data.
This work presents TrustFlip, a novel attack that weaponizes consistency-based defenses to poison the trust assigned to benign vehicles, and introduces TrustReflect, a lightweight self-reflection mechanism that marks disputed regions as uncertain and excludes them from trust evaluation, reducing the attack success rate...
It is argued that evaluation should move beyond static question answering and model accuracy towards workflow-level assessment of evidence quality, tool use, policy and invariant compliance, staged execution, recovery, calibration, cost, and human intervention.
This work introduces a deterministic full-chain memorization mechanism that locks onto token-level secrets in dynamic computation flows via online tensor-rule matching, and leverages value-gradient decoupling to stealthily inject attack gradients, overcoming gradient drowning to force model memorization.
Reach audiences
Advertise in front of researchers, engineers, and readers.
AgenTRIM is introduced, a framework for detecting and mitigating tool-driven agency risks without altering an agent's internal reasoning that provides a practical, capability-preserving approach to safer tool use in LLM-based agents.
JAWS-BENCH(Jailbreaks Across WorkSpaces), a benchmark spanning three escalating workspace regimes mirroring attacker capability, is presented, indicating that JAWS-BENCH can be reused across multiple agent frameworks.
NeuroBreak is presented, a visual analytics system that helps experts progressively unpack jailbreak mechanisms from layer-level semantics down to neuron-level behaviors and provides actionable insights for strengthening LLM defenses.
It is proved that, under an intuitive and experimentally supported assumption called distribution consistency, obfuscation can nullify the robustness of N-gram-based watermarks and motivate more semantics-aware alternatives.
The first Fraud Investigation Assistant (FIA) framework is introduced, which employs multimodal large language models (LLMs) to automate key steps of credit card fraud investigation and generate explanatory reports and suggests that LLM-based agents can assist with automating substantial parts of the fraud investigatio...
REP, a lightweight in-context elicitation method that uses shadow-model-generated demonstrations wrapped in auxiliary code-like formats to raise user-visible reasoning traces from a victim model, substantially increases similarity between exposed and REP-conditioned internal traces while preserving useful reasoning sig...
A four-stage forensic audit protocol for API-served models, validated prospectively on a flagship case whose 2026-08-23 analysis pointed to the GLM-5.3 version line and whose official reveal confirmed those family and version-line inferences.
SIR is presented, a black box IPI attack that composes stealthy injections from a small library of reusable principles stated in plain language and wraps composition in an iterative feedback loop that diagnoses the victim's failed trajectories and distills the bypasses into new, named strategies that are reapplied acro...
Research Manager Mike Walker reflects on a career spanning hacker culture, cybersecurity, and AI, sharing lessons about trust, teamwork, leadership, and why solving difficult problems requires curiosity, judgment, agility—and an understanding of the rules. The post Learn the rules: The hacker mindset appeared first on Microsoft Research.
GPT-Lab
· gpt-lab.eu
Oct 30, 2026
Critical flaws, compromised release channels and agentic AI threats landed within days of each other in mid-October 2026. Why “Black October” may say more about how well we now find weaknesses than about the attacks themselves. The post Black October? The Week AI Changed the Cybersecurity Conversation appeared first on GPT-Lab.
Gemini 4 Carbon is our strongest model with state-of-the-art cybersecurity and long horizon capabilities.
Gemini 3.8 Flash and 3.8 Flash Cyber deliver next-generation intelligence for agentic workflows and cybersecurity.
Google introduces Gemini 3.5 Flash Cyber, a lightweight cybersecurity model to find and patch vulnerabilities.
Students from the MIT Cybersecurity Clinic help local governments and other vulnerable organizations defend against digital threats.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
