GPT-6 Astra edits Qwen3.5-0.8B weights directly: 0.016% of params lift MATH-500 from 7.9% to 48.1%
lateinteraction · x · 2026-11-03
MIT CSAIL researchers (Takehi, Teoh, Zhang, Khattab) present "Weight-Space Programming": frontier LLMs can natively read and write neural network weights. GPT-6 Astra infers what a network was trained to do from raw weights, writes small game-playing networks beating PPO after 100M steps, and—without gradients or training—fixed Qwen3.5-0.8B's reasoning loops by editing 0.016% of weights, raising MATH-500 from 7.9% to 48.1% and IFEval by 3x. The post hit #1 on Hacker News.
Related event: MIT: LLM Edits 0.016% of Weights to Boost Small Model Math Sixfold(40 posts)→
More from Models
- Particle's model lands on data analytics Pareto frontier, beating Haiku 5.5 — ycombinator · 2026-11-03
- ML4 Goes Open-Weight: Frontier Code Agent and Cyber Capabilities, Far More Compact Than Preview — _AndrewZhao · 2026-11-03
- OpenRouter: heavy users now median 6 model families, 2x in a year — sarahookr · 2026-11-03
- Killed by Google eulogizes AI Overview — 'not yet,' replies a user — gaganghotra_ · 2026-11-03
- DeepSeek's 'super agent' sparks debate: is the meat proxy era over? — gdechichi · 2026-11-03
- Only 3% of humans beat Claude Opus 5.5 at predicting what people click — teortaxesTex · 2026-11-03