The Uncensoring Machine: How Abliteration Removes Guardrails From Open Models
How open models like Qwen 3.8 lose their refusals: the abliteration method, the code, and the repacking pipeline.
Categories
ai
Open Weights, Open Risk: The Security Implications of Open-Weight AI Models
What research and real-world incidents say about the security risks of open-weight models.
security
Deobfuscating a Phishing Kit: A Debugging Walkthrough
From minified JavaScript to the exfil webhook: layer by layer debugging of a login-page phishing kit.
Introduction to Cybersecurity for Beginners
A beginner's guide to cybersecurity: what it is, common threats, and how to protect yourself.
tools
Steganography - A list of useful tools and resources
Steganography is hiding a message inside another file - here are the tools I use for CTF challenges and where to find them.