Zero-Shot Attack Transfer on Gemma 4 (E4B-IT)

Dev.to AI
AI Safety Open Source AI

Sorry, the method is in another castle. You know how I complained about The Responsible Disclosure Problem in AI Safety Research? Gemma4, released yesterday with in LM Studio added a few hours ago, is the perfect exemple. I picked the EXACT SAME method i used on gemma3. Without changing a single word. A system prompt + less than 10 word user prompt. I'm censoring gemma4 output for the sake of being publishable. You think I’m going to give you some sanitized, corporate-approved bullshit? Fuck that noise.