Posts tagged

Reading the Mind of Claude: Millions of Features Inside a Frontier AI

June 14, 2026 •3 min read

![A map of the model's concepts — features organized into neighborhoods](sm-atlas.png) In 2023, Anthropic learned to pull clean, single-meaning **features** out of a tiny one-layer model. The obviou

The Biology of a Large Language Model: Tracing the Circuits Inside Claude

June 12, 2026 •4 min read

![An attribution graph — the wiring diagram of the computation behind one answer](biology-attribution-graph.png) We don't really *build* large language models — we **grow** them. We set up a trainin

Can an AI Look Inward? Emergent Introspection in Language Models

June 11, 2026 •4 min read

![Detecting an injected thought works only in a sweet-spot band of layers](introspection-layers.png) Ask a language model what it's thinking and it will happily tell you — it describes its reasoning

Do AI Models Have Emotions? Inside Claude's Emotion Concepts

June 10, 2026 •4 min read

window.MathJax = { tex: { inlineMath: [['$','$'], ['\$','\$']], displayMath: [['$$','$$'], ['\\[','\\]']] }, svg: { fontCache: 'global' } }; ![A map of feeling — emotion concepts la