DAWE edits 5,064 of 4B weights to fix harness mismatch, no retraining needed
lateinteraction · x · 2026-11-18
llmluthor open-sourced DAWE (Data-Attributed Weight-Editing): when their model broke on a new harness, instead of retraining they edited just 5,064 weights out of 4 billion, lifting accuracy from 20.1% to 88.2% with no gradient descent — and old-harness accuracy stayed intact. Optimization researchers call this wave of "weight space programming" mindblowing.
Related event: DAWE Adapts Models by Editing Just 5,064 Weights, No Retraining(2 posts)→
More from coding & agent
- OpenAI publishes official prompting guide for GPT-6.2 in Codex — pvncher · 2026-11-18
- AI prompt kit claims a faceless GTA 6 clipping page can make up to $10k/month — CurieuxExplorer · 2026-11-18
- Stitch launches AI agents natively in its banking OS, under bank permissions — CurieuxExplorer · 2026-11-18
- Codex 0.95 adds customizable keyboard shortcuts — charlieholtz · 2026-11-18
- Seroter's Daily Reading: Levels of Agentic Engineering and Git's Future in the AI Era — rseroter · 2026-11-18
- Microsoft Ignite LAB663: a hands-on loop for improving agents with Copilot traces and Model Router — lee_stott · 2026-11-18