C
Sep 15, 2026Context-Augmented KG Training Alone Wont Make LLMs Safe for Professional Use
Canonical version:...
Sep 15, 20267 min read0 reactions0 comments
Back to articles 
Apr 25, 2025
Tag archive
#llmsafety
T
Aug 14, 2026Tool-Permission Gate: The Only MiniMax H3 Test That Matters
When MiniMax H3 starts trending, the question I ask is not ‘What did it score on MMLU?’ but ‘Will it...
Aug 14, 20265 min read0 reactions0 comments
V
Aug 14, 2026VIDRAFT AX-Ray: Detecting Causal-Leakage Vulnerabilities in General-Purpose LLMs — Leaderboard Now on Hugging Face
VIDRAFT가 공개한 AX-Ray는 LLM의 인과누설(Causal Leakage) 취약점을 진단하는 AI 안전성 평가 프레임워크. 117개 진단 항목과 다국 법규 매핑, Hugging Face 리더보드 제공.
Aug 14, 20264 min read0 reactions0 comments
W
May 16, 2026What Is AI Jailbreaking? The Security Challenge Reshaping LLMs
What Is AI Jailbreaking? The Security Challenge Reshaping LLMs The term jailbreaking has...
May 16, 20263 min read0 reactions0 comments
I
Mar 31, 2026Identity + Governance = 100% Safety? Testing Combined Persona Approaches on Abliterated LLMs
In our previous experiment, we showed that persona-level behavioral rules (Soul Spec) barely help...
Mar 31, 20265 min read0 reactions0 comments
Type Guards in TypeScript 2025: Next-Level Type Safety for AI-Era Developers
By 2025, 47% of codebases use AI tools (GitHub 2025 Insights), but TypeScript remains the #1 defense...
Apr 25, 20252 min read2 reactions2 comments