Job Mendes·Sep 13AI Security Learning Journal — Day 11Jailbreaking LLMs: When Safety Is a Probability, Not a Wall
InfosecTrain·Aug 20How Do Jailbreaking Attacks Exploit Large Language Models (LLMs)?Large Language Models (LLMs) are programmed to refuse dangerous or harmful requests. Developers use techniques like Reinforcement Learning…
Cory Doctorow·Aug 14Capital formationGoing legit means going mainstream.A response icon7A response icon7
Khanmughees·Aug 11Jailbreaking DeepSeek into Building a Keylogger — Then Hunting It DownEducational lab experiment: LLM red teaming, malware analysis, and defensive detection.
Mati OS·Jul 24Jailbreaking de LLMs mediante Prompt InjectionJailbreaking mediante Prompt Injection y Prompt Engineering
YouShin kim·Jul 20[Google DeepMind / LASR Labs] — 투명성의 배신: AI의 ‘생각의 흐름(CoT)’ 감시망이 탈옥 공격에 무너지는 이유1. 한눈에 보는 핵심 요약 (Executive Summary)
Arvind Kumar·Jul 2Jailbreaking LLMs: How Attackers Bypass AI Safety Controls and What Engineers Must Do About ItThe AI Security Problem Nobody Talks About Enough
thatqagirl·Jun 30🛡️ AI Guardrails Are Becoming the First Line of QualityAs AI systems become part of production applications, testing alone is no longer enough.
The broken General of Techetonium·Jun 28Inside the Guardrails: Why AI Safety Filters Exist and Where They Actually FailEvery time a screenshot of a ‘jailbroken’ model goes viral, the public conversation collapses into one of two bad takes: either ‘the…
![[Google DeepMind / LASR Labs] — 투명성의 배신: AI의 ‘생각의 흐름(CoT)’ 감시망이 탈옥 공격에 무너지는 이유](https://webproxy.poorya-velaei-d67.workers.dev/https://miro.medium.com/v2/resize:fill:160:106/1*rT3OZdP04U3xvNSR36B6rg.png)
![[Google DeepMind / LASR Labs] — 투명성의 배신: AI의 ‘생각의 흐름(CoT)’ 감시망이 탈옥 공격에 무너지는 이유](https://webproxy.poorya-velaei-d67.workers.dev/https://miro.medium.com/v2/resize:fill:320:214/1*rT3OZdP04U3xvNSR36B6rg.png)
