GPT-6 edits Qwen3.5 weights directly: 0.016% of weights changed, math improves up to 6x
lateinteraction · x · 2026-11-03
rikiyatakehi's team demonstrates "Weight-Space Programming": GPT-6 Astra directly edited Qwen3.5-0.8B weights to improve math, changing only 0.016% of weights with no gradients or training. It fixed the small model's reasoning loops, improved math up to 6x and instruction following 3x, suggesting frontier LLMs can natively read and write neural network weights.
More from Models
- Inference providers can't survive without DeepSeek-grade KV-cache engineering, argues thread — AccBalanced · 2026-11-03
- Fabricated citations collapse from 55% to 0%: new models nearly fix the hallucinated-reference failure — RexDouglass · 2026-11-03
- Apple hints at monetizing third-party AI apps for new revenue on earnings call — TechCrunch AI · 2026-11-03
- Writing Weights Index: Open-Source Models Struggle at Weight-Space Programming — eliebakouch · 2026-11-03
- Gemini 4 Argon ships, putting Google back in the frontier model race after a year-long lull — danshipper · 2026-11-03
- Hands-on audit punctures Astra's '0.016% weight programming' claim: it just disabled Qwen's thinking mode — teortaxesTex · 2026-11-03