Locking it down: A new technique to prevent LLM jailbreaks
Following on from our preview, here’s the full rundown on LLM salting: a novel countermeasure against LLM jailbreaks, developed by AI researchers at Sophos X-Ops
Similar Posts
AI Slashes Workloads for vCISOs by 68% as SMBs Demand More – New Report Reveals
Bys sAs the volume and sophistication of cyber threats and risks grow, cybersecurity has become mission-critical for businesses of all sizes. To address this shift, SMBs have been urgently turning to vCISO services to keep up with escalating threats and compliance demands. A recent report by Cynomi has found that a full 79% of MSPs and…
Citrix Bleed 2 Flaw Enables Token Theft; SAP GUI Flaws Risk Sensitive Data Exposure
Bys sCybersecurity researchers have detailed two now-patched security flaws in SAP Graphical User Interface (GUI) for Windows and Java that, if successfully exploited, could have enabled attackers to access sensitive information under certain conditions. The vulnerabilities, tracked as CVE-2025-0055 and CVE-2025-0056 (CVSS scores: 6.0), were patched by SAP as part of its monthly updates for January
Google Pays $1.375 Billion to Texas Over Unauthorized Tracking and Biometric Data Collection
Bys sGoogle has agreed to pay the U.S. state of Texas nearly $1.4 billion to settle two lawsuits that accused the company of tracking users’ personal location and maintaining their facial recognition data without consent. The $1.375 billion payment dwarfs the fines the tech giant has paid to settle similar lawsuits brought by other U.S. states….
Have You Turned Off Your Virtual Oven?
Bys sYou check that the windows are shut before leaving home. Return to the kitchen to verify that the oven and stove were definitely turned off. Maybe even circle back again to confirm the front door was properly closed. These automatic safety checks give you peace of mind because you know the unlikely but potentially dangerous…
CISA Adds Four Actively Exploited Vulnerabilities to KEV Catalog, Urges Fixes by Feb 25
Bys sThe U.S. Cybersecurity and Infrastructure Security Agency (CISA) on Tuesday added four security flaws to its Known Exploited Vulnerabilities (KEV) catalog, citing evidence of active exploitation in the wild. The list of vulnerabilities is as follows – CVE-2024-45195 (CVSS score: 7.5/9.8) – A forced browsing vulnerability in Apache OFBiz that allows a remote attacker to…
AI-Powered Villager Pen Testing Tool Hits 11,000 PyPI Downloads Amid Abuse Concerns
Bys sA new artificial intelligence (AI)-powered penetration testing tool linked to a China-based company has attracted nearly 11,000 downloads on the Python Package Index (PyPI) repository, raising concerns that it could be repurposed by cybercriminals for malicious purposes. Dubbed Villager, the framework is assessed to be the work of Cyberspike, which has positioned the tools as…