yoavgo: inferring network behavior from weights impresses him; hand-writing weights is just scaled-up exam tasks
yoavgo · x · 2026-11-03
Yoav Goldberg (yoavgo) says he's mostly impressed by experiments that interpret what a network does based on its weights. The "writing weights from scratch" demos are far less impressive to him — he's had students write network weights for tasks in exams, and this simply scales that up, which is perfect for a computer.
Related event: Yoav Goldberg: reading weights from scratch(2 posts)→
More from Models
- AI writes Breakout network weights in one reply, scoring 20× above PPO trained 500M steps — xennygrimmato_ · 2026-11-03
- DeepSeek's Underrated Role: Scaling and De-risking Ideas Academia Published Years Earlier — teortaxesTex · 2026-11-03
- Do faster quantized LLMs (Basalt, Infernix, Strata-nvfp4) trade away quality? — Kooky-Mode3047 · 2026-11-03
- OpenAI gives DevDay attendees 30 days of free Pro 100 tier — testingcatalog · 2026-11-03
- Peter Yang: Gemini beats ChatGPT at copy editing, but its harness holds it back — petergyang · 2026-11-03
- Mistral publishes Le Chonk technical report, praised as a rare substantive LLM writeup — _AndrewZhao · 2026-11-03