STATUS LOG
// NOTICE: This blog space is currently undergoing architectural and content-focused upgrades. Some articles may temporarily remain as offline drafts.
Latest Publications
AI Safety for Noobs (Public Education)
How I Broke My Own Image Classifier (And What It Taught Me About AI Safety)
I changed exactly one pixel in an image of a "3" — invisible to the human eye — and my neural network confidently declared it was an "8".
Apr 2026
AI Safety for Noobs (Public Education)
Can Mechanistic Interpretability Detect Who Is Asking? (Elderly or Child)
Stop treating models like black boxes. Learn how to map, audit, and steer internal demographic circuits to identify if a child or an elderly person is asking the question.
AI Safety for Noobs (Public Education)
[DRAFT LOG]
Linguistic Roots of AI Safety: An Entry Analysis of LLM Value Alignment
Investigating standard value metrics across diverse regional representations and translation constraints.
Reflections
[DRAFT LOG]
AI Safety: Alignment – After 10 Weeks
Reflections on deep system exploration, training mechanics, and alignment benchmarks.
AI Governance
[DRAFT LOG]
Reflection on My AI Governance Application
A journey mapping national frameworks, safety guidelines, and regulatory priorities.
Peta Penyelidikan // Wacana Kognitif & Pengurusan Sumber Pengiraan