# Tree-Ising: Repairing Language Models with Bit-Level Ising Edits

> Source: <https://aiflash.com/news/149591/>
> Published: 2026-11-11 05:30:09+00:00

Weight-space programming, introduced in recent MIT work by Takehi et al., enables a frontier LLM agent to read a network's weights and write a repair. Their GPT-6 Astra agent corrected the reasoning loops of Qwen3.5-0.8B without gradients, using 158 agent turns, 89.7 million tokens, and 50 forward-p