GPT-6 Astra can edit Qwen3.5-0.8B’s weights and raise its MATH-500 from 7.9% to 48.1%,…
This is a dev post classified by AGK Intelligence as AI dev tools (a free resource), kept by the Dev Radar because it carries real work, not commentary.
GPT-6 Astra can edit Qwen3.5-0.8B’s weights and raise its MATH-500 from 7.9% to 48.1%, without gradients or retraining. “Sparks of Weight-Space Programming” This blog from MIT asks whether LLMs can read and edit neural network weights directly, much like code. In Qwen, Astra traced repetitive reasoning loops to a direction written by two MLP layers, then edited the weights to suppress it. With 50 batches of forward-pass feedback, the fix also raised instruction following from 21.0% to 62.3% Read on alphaXiv: https://www.alphaxiv.org/abs/2611.weight-space-programming
Posted by alphaXiv (59.9k followers) 2 h ago · 80 likes · 3.5k views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- Too good of a result to not share, DeepSeek-v4.1-Flash-1017 ofcourse ! — @ItsmeAjayKV
- you cannot be serious... new weight writing bench just dropped — @zainhas
- It’s extremely important that the open sourced community is made aware of all… — @ViC305
- I benchmarked 5 computer-use tools — @athrix_codes
- Post novo: "Criando um framework para IA com IA: REA Launcher SDK" — @AkitaOnRails
- Test AI trading agents without putting real money on the line. — @DanKornas
- 🚨Pulse Alpha is almost certainly a new model from Thinking Machines, and it's already… — @SPAC89
- Claude wrote entire working transformers by hand. — @0x0SojalSec
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.5k posts from 5k X accounts over the last 21 days, 9.1k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-03 02:32 UTC. Full method.