The name is the button.

An agent writes the code now, and a human thumb hovers over Accept all a hundred times a day. The honest question of this decade is what, exactly, you just accepted. So I read the diff.

Eight years as a backend engineer in Stockholm, back in Tokyo since 2023, and I still ship something small most nights, usually badly. My one rule: the only real review of a model is the thing you built with it. That is why there is no leaderboard on this site.

Around a hundred and forty sources come through — the labs' own posts, the papers, the tools, the runtimes and languages nobody is watching yet, the chip under the demo, and the newsletters people actually forward. What is moving gets counted so the week reads at a glance, and every few days I explain one mechanism in plain words.

A story here is a headline, a short excerpt, one image and the link to whoever wrote it. We never reproduce the post; the images belong to their authors. A post written in Japanese appears in Japanese, exactly as it was published, and the link goes to the original.

— Yuki Halvorsen, Tokyo

← Accept All Archive

Transformer CircuitsResearch

Characterizing interference weights in a tiny language model

08.21

We identify interference weights in a 1-layer transformer by measuring their effect on model outputs and loss.

Models & releases note

Read at Transformer Circuits ↗

One letter a week, starting soon.

Latest

Characterizing interference weights in a tiny language model

Related

ResearchEmotion Concepts and their Function in a Large Language Model Transformer Circuits

ToolsGLM-5.3 vs. GPT-5.6 Sol on DeepSWE: Cost, Coding, and Routing Together AI

ToolsGLM-5.3 vs. Claude Fable 5 on DeepSWE: Cost, Coding, and Routing Together AI

ToolsOpen-sourcing evals for open-weight agents Cline Blog

ToolsDeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and Routing Together AI

ToolsHow we rebuilt the Auggie CLI harness to make tasks 53% cheaper than Claude Code Augment Code