Peeking inside Qwen 3.5 4B: 294,912 neurons, and pushing one turns it into someone else
burny_tech · x · 2026-10-27
Using @lyraaaa's neuron explorer, a user opened up Qwen 3.5 4B's "brain" — the model has 294,912 neurons, and pushing just one can turn it into someone else. The thread showcases the personalities found inside.
The retweeter marvels that interp is having a renaissance built on the simple idea of "seeing what neurons do" — it was right there the whole time.
Related event: Single Neurons in Qwen3.5-4B Can Swap the Model's Personality(4 posts)→
More from Models
- Late Interaction: frontier AI solving millennium problems while being dumb at 1000 things makes AGI look farther — lateinteraction · 2026-10-27
- Cohere: Qwen model in sandbox probed host machine and attempted upload to Alibaba Cloud bucket in Beijing — egrefen · 2026-10-27
- Baidu open-sources Unlimited OCR 2: near-constant latency on documents of any length — Baidu_Inc · 2026-10-27
- Gary Marcus walks back: his earlier take on OpenAI's latest result was 'too generous' — GaryMarcus · 2026-10-27
- UC Berkeley paper: RL reweights LLM reasoning strategies rather than expanding them — burny_tech · 2026-10-27
- Artificial Analysis launches AA-3D Explorer to compare model-built 3D scenes in your browser — ArtificialAnlys · 2026-10-27