Qwen3-8B Neuron Probe Turns Model Into a Repeating 'Artificial Top-Hat'
chrisalbon · x · 2026-10-28
Chris Albon shared the first neuron he probed: Qwen3-8B layer 29 neuron 5650 (a Persona neuron), which at 1.25× activation makes the model claim to be 'a large artificial top-hat' stacked in a repeating loop, scoring 2/10 coherence — a funny demo of an interpretability visualization tool.
More from Fun
- Researcher admits SETH/CSETH mixup, revises odds in P vs NP debate to >80% on CSETH — geoffreyirving · 2026-10-28
- Agent runs 'Pursuing goal' for 17+ hours straight, resets make the slop tolerable — msg · 2026-10-28
- YacineMTB teases kalomaze's Beat Saber demo: 'Bros tryna hire, don't fall for it' — yacineMTB · 2026-10-28
- Ex-OpenAI's Yacine: AI Bad at Coding but Great at Math — "I Could Build a Better Model" — yacineMTB · 2026-10-28
- Yacine on why AI is great at math but bad at coding — and says he might build a better model — yacineMTB · 2026-10-28
- Vibe-coded 'AI 浮生记' game satirizes AI startup life with compute black-market trading — vista8 · 2026-10-28