Welcome to Nokuno AI — a new blog about machine learning research. The goal is simple: write down the parts of ML that deserve more attention than a 140-character thread — close readings of papers, honest notes on why methods fail, and practical guides for doing research with large models.

Coverage will center on a few themes the author works on and cares about: interpretability (sparse autoencoders, circuits), large language models (reasoning, multimodal behavior, evaluation), robustness, and the research process itself.

New posts will appear on the blog page and under Research Notes. Feedback, corrections, and paper tips are welcome — that’s what the contact form is for.

— About Nokuno AI