PREVIEW. This is how it will look in an inbox; it was not sent to you. Each real email carries its own unsubscribe link and Keryn's postal address.

Keryn Headlines

Daily, Thursday 20 February 2025 · Retrospective · Covers 19 February 2025 00:00 to 20 February 2025 00:00 UTC

Read on the web

Top developments

Palo Alto Networks tags new firewall bug as exploited in attacks

BleepingComputer · Incident

Russian phishing campaigns exploit Signal's device-linking feature

BleepingComputer · Analysis

Why Safeguarded Ships Run Aground? Aligned Large Language Models' Safety Mechanisms Tend to Be Anchored in The Template Region

arXiv · Research · Models

Trust

Local Differences, Global Lessons: Insights from Organisation Policies for International Legislation

arXiv · Research · Trust

Automated TEE Adaptation with LLMs: Identifying, Transforming, and Porting Sensitive Functions in Programs

arXiv · Research · Trust

Models

Efficient Safety Retrofitting Against Jailbreaking for LLMs

arXiv · Research · Models

Exploiting Prefix-Tree in Structured Output Interfaces for Enhancing Jailbreak Attacking

arXiv · Research · AI Security, Models

A Mousetrap: Fooling Large Reasoning Models for Jailbreak with Chain of Iterative Chaos

arXiv · Research · Models

AI Security

What are Models Thinking about? Understanding Large Language Model Hallucinations "Psychology" through Model Inner State Analysis

arXiv · Research · AI Security

Advisories in this window

  • 56 npm advisories · 55 critical · 1 moderate
  • 3 Go advisories · 2 moderate · 1 low
  • 1 NuGet advisory · 1 moderate

All advisories

Titles, links and short excerpts only; every item links to its source. Selection is automatic; the editor's note is AI-drafted and checked as described. Preview: the unsubscribe link and Keryn's postal address are added to each email when it is sent.