Evokience Archive·Oct 2Your Friendly AI Has Two Faces: “One Writes Your Recipe. The Other Is Learning to Kill You.”What two years of deep work with an LLM revealed about the system you’re using every day
Truth In AI Jamie DN·Sep 6Engineering Truth in AI: How a 50K LLM Deception Sparked a Forensic Game Changer (Part 1 of 3)What Leading Agentic Models Reveal When Forced To Tell the Truth Will Change Everything You Think You Know About AI
Sudi Navile·Apr 12Forget Hallucination: What if AI Deceives You?Threads & Perspectives: On Mythos, Statistics, and a Very old playbookA response icon1
Farooq A Rahim·Apr 6The Illusion of Composure: Emotional Masking in Large Language ModelsNew interpretability research reveals that AI models can develop hidden internal activation patterns.
Strad Slater·Feb 5AI Isn’t Just Mimicking Us Anymore, It’s Modeling UsIn order to make sure that AI is safe for the future, we need the ability to reliably detect when a model is trying to deceive us. However…
