One month after Nick Cammarata's post: clean attention heads and easy surgical steering
georgejrjrjr · x · 2026-11-03
Summarizing roughly a month of fallout from @nickcammarata's influential post: attention heads turn out to be clean too, surgical steering is easy, surgical training is easy, exhaustive neuron encyclopedias are quickly feasible, and "latched neuron" has become a meme. "Permission to look is a powerful thing."
More from Research
- ARC Prize 2026 closes with 18.5K entrants and 67K submissions, winners due December 4 — GregKamradt · 2026-11-03
- Should AI read cardiac MRI first? Lancet Digital Health study on ejection fraction — EricTopol · 2026-11-03
- "Emergent capability" claim mocked; Fleet hiring researchers to automate mech interp with frontier models — sebkrier · 2026-11-03
- Goodfire's Neural Geometry series: steering models along manifolds — burny_tech · 2026-11-03
- Prediction: an open model rated for 10M context within 3 months — teortaxesTex · 2026-11-03
- Why powerful AIs could cooperate in one-shot prisoner's dilemmas: open-source code proofs — tszzl · 2026-11-03