Looped transformer analysis pins down single neurons for non-duality, phonetics, and other concepts
burny_tech · x · 2026-10-27
Analysis of a looped transformer identified five concept-bearing single neurons: phonetics/letters (L18 N3066−), non-duality (L13 N1680+), pedagogy (L4 N2104+), forum "accepted answer" (L6 N3344−), and electron orbitals (L18 N1735−). Clamping L18 N3066− makes a news sample veer into IPA-style phonetics talk, showing single neurons can steer output toward a concept domain. Follows the author's earlier "nonduality neuron" find.
More from Research
- Claude Opus 5.5 claims ETH false, all-dimension Kakeya and Yang-Mills in one day — repligate · 2026-10-27
- Anthropic internal Claude model claims 2^{o(n)} 3-SAT algorithm refuting ETH — imjustnewatai · 2026-10-27
- Companion robot Moflin learns to comfort via local model and JEPA world model — Stefania_druga · 2026-10-27
- Complexity researcher on the ETH refutation: stronger than P≠NP but practically moot — burny_tech · 2026-10-27
- Complexity theorists to independently examine OpenAI model's claimed ETH refutation — burny_tech · 2026-10-27
- Steering Just 4 Neurons Makes a Model Spell Out an AI Domination Plan — burny_tech · 2026-10-27