Researcher shows OpenAI's GPT-6 Astra hand-editing Qwen's weights without training

Takehi publishes "Sparks of Weight-Space Programming" findings

1 Nov 2 1:21 PM · 6d ago · 2 articles · 1 post · 2 sources · development 1 of 5

The post shows GPT-6 Astra inferring a network's trained purpose from raw weights, writing game-playing networks from scratch that beat PPO-trained ones, and editing Qwen3.5-0.8B's weights to fix a reasoning-loop bug, raising MATH-500 from 7.9% to 48.1% and IFEval 3x.

Takehi publishes "Sparks of Weight-Space Programming" findings — rikiyat.github.io

“Just by looking at the raw weights, GPT-6 Astra can tell what a network was trained to do.”

Rikiya Takehi

Rikiya Takehi Researcher, author of the blog postGPT-6 Astra OpenAI LLM used as the weight-editing modelQwen3.5-0.8B Target model edited by Astra

The whole story articlespostscomments the bright band is this development · numbered dots are the others · click one to jump

1–5Takehi publishes "Sparks of Weight-Space…

Nov 3Nov 4Nov 5Nov 6yesterdaynow · 1:51 PM ET

What was reported 1 claim about this development

2 outlets Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights

first by Lobsters, 5d ago · also Rikiya Takehi
1 more headline
- Researchers say LLMs can be good at natively reading and writing neural network weights; GPT-6 Astra edited Qwen3.5-0.8B's weights to make it better at math Rikiya Takehi · 5d ago

All 5 developments of Researcher shows OpenAI's GPT-6 Astra hand-editing Qwen's… →

MastodonHacker NewsNewswiresLobstersXRedditBluesky