MMartin Bowling·Oct 27The Celebrity Neuron Is Not a NeuronI tried to make Qwen3.5-4B identify as Einstein, Curie, Newton, Taylor Swift, and Keanu Reeves. The scientists worked. The entertainers didn’t.5 min read·martinbowling.com
MMartin Bowling·Oct 26I Audited the Viral Qwen Neuron ZooI tested the viral one-neuron personas on Qwen3.5-4B. 16 of 22 survived a one-digit fix, one was impossible, and three neurons become the Golden Gate Bridge.10 min read·martinbowling.com
MMartin Bowling·Oct 26Three-Neuron BangersFive combos that fuse into something none of the neurons does alone — plus two cursed loops and one random control that does nothing.5 min read·martinbowling.com
EEkansh Nangia·Sep 17Introduction To Mechanistic InterpretabilityHow do LLMs actually work? Explore mechanistic interpretability, AI epistemology, and why probing neural networks cannot crack the AI black box.18 min read·ekanshn.hashnode.dev
RReid Marlow·Jul 11Anthropic Found the Hidden Space Where Claude Thinks. It's Weirder Than You'd Think.Anthropic Found the Hidden Space Where Claude Thinks. It's Weirder Than You'd Think. Anthropic just dropped a paper that gave us the clearest window yet into what an LLM is doing between reading your6 min read·reidmarlow.hashnode.dev1
RRaghavan T M·Jul 2Model Limitations: The Two Everyone Knows and the Ones Nobody Mentions Beyond hallucinations: Discover the hidden limits of AI models, from sycophancy and the reversal curse to the deep interpretability gap.6 min read·changeofbasis.hashnode.dev
YYejun Park·May 11From One Tree to a Forest: A Practical Guide to Ensemble Methods in Machine LearningTurn one fragile tree into a thriving forest: practical ensemble techniques that stabilize and boost performance14 min read·yejunpark3.hashnode.dev
CClawSouls·Apr 5Anthropic Proved AI Has Functional Emotions — Persona Design Is Now a Safety IssueThey Looked Inside the Brain Anthropic's Interpretability team just did something unprecedented. They opened up Claude Sonnet 4.5's neural network, mapped 171 emotion concepts to specific patterns of artificial neurons, and proved these patterns dire...5 min read·clawsouls.hashnode.dev
MManas Vardhan·Mar 21Your LLM Is Thinking More Than It Tells YouTwo papers explain why that's a problem - and reveal that reasoning makes models more honest, not less.8 min read·theagentstack.hashnode.dev
CClawSouls·Feb 28What Anthropic Just Proved — AI Personas Aren't Prompts, They're IdentitiesAnthropic's newly published Persona Selection Model research answers the most important question about AI personas: AI behaves like a human not because it's programmed to, but because it's an inevitable consequence of learning. The Core Finding: AI P...4 min read·clawsouls.hashnode.dev