Qwen 3.8 27B on my laptop can interpret model weights, too Simon Willison reported that Qwen 3.8 27B, running locally on his laptop, could identify which game a small reinforcement-learning policy network was trained to play by reading the network's raw weights, replicating an example from the paper "Sparks of Weight-Space Programming" by Rikiya Takehi, Jayden Teoh, Alex Zhang and Omar Khattab. That paper reported that GPT-6 Astra could read and edit model weights directly, including fixing Qwen3.5-0.8B's reasoning loops by editing 0.016% of its weights, which raised MATH-500 from 7.9% to 48.1% and tripled IFEval. The claims drew criticism from spaCy author Matthew Honnibal, who called the paper misleading for conflating tool-free weight reading with code-assisted weight updates, and from Sean Goedecke, who argued the Qwen3.5-0.8B example followed steps a human ML researcher would take. Yesterday's hot paper was Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights https://rikiyat.github.io/blog/weight-space-programming/ by Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab. The paper found that frontier models, in particular GPT-6 Astra, can both interpret and modify the raw weights of a machine learning model - literally a text description of the model containing a bunch of floating point numbers. I've been playing with one of the examples from the paper and I'm delighted to share that Qwen 3.8 27B running on my laptop can do this as well Here's the highlight from their abstract, describing impressive capabilities for GPT-6 Astra and coining the term Weight-Space Programming for when LLMs analyze and modify model weights: Just by looking at the raw weights, GPT-6 Astra can tell what a network was trained to do . And without any environment feedback, it has no trouble writing, from scratch, small game-playing networks that outplay networks trained with PPO for 100 million steps. Likewise, asked to make Qwen3.5-0.8B better at math without gradients, Astra identifies an issue in its weights causing reasoning loops that Qwen’s own model card https://huggingface.co/Qwen/Qwen3.5-0.8B/blob/main/README.md L856 warns about, and fixes this by natively editing 0.016% of the weights . With the same sampling settings, this weight correction raises MATH-500 from 7.9% to 48.1% and generalizes out of the box to instruction following, raising IFEval by 3x. All of this is possible without allowing Astra to write training loops. It uses its reasoning to figure out what each part of a neural network computes, or should compute, and changes the weights accordingly. We call this Weight-Space Programming . The paper combines two topics that I think deserve to be treated separately. The first is what models can do with raw weights without running tools or executing code. The second explores how well models, once armed with code, can run operations to update and improve other models. spaCy author Matthew Honnibal called the paper misleading https://honnibal.dev/blog/astra-read-write-misleading for conflating the two, and Sean Goedecke deconstructed the Qwen3.5-0.8B update example https://www.seangoedecke.com/llms-are-not-superhuman-at-reading-and-writing-neural-network-weights/ as taking the same steps that a human ML researcher would use. I'm going to focus on the no-tools question, which I think is the more interesting of the two. The paper frames this as Can LLMs understand model behaviors from weights? They trained four neural networks to play simplified 10x10 grid versions of four classic computer games - Breakout, Pong, Pac-Man, and Freeway I had to look that one up