# Simon Willison — Qwen 3.8 27B reads model weights on a laptop

> A local 27B model can do a slow version of the weight-reading task from the GPT-6 Astra paper.

- Org: Simon Willison
- Date: 2026-11-03
- Category: article
- Source: https://simonwillison.net/2026/Nov/3/qwen-can-interpret-model-weights/
- Canonical page: https://ai-tldr.dev/releases/simon-willison-qwen-interprets-model-weights/

## In short

Simon Willison ran Qwen 3.8 27B locally and it named the game behind three of four policy networks from raw weights. It failed on Freeway. Kimi K3.1 Pro got Breakout right in about 10 minutes.

## What it is

Simon Willison tested whether a local model can reproduce the no-tools weight-reading task from the paper "Sparks of Weight-Space Programming". The task is to name which of four games (Breakout, Pong, Pac-Man, Freeway) a policy network was trained on, using only its weights as text.

## How it works

He ran a 17GB Q4_K_M build of Qwen 3.8 27B in LM Studio through the llm-lmstudio plugin. Qwen named Pac-Man, Pong and Breakout with 72% to 88% confidence, taking 46 to 65 minutes each, and got Freeway wrong. Kimi K3.1 Pro, run via OpenRouter, identified Breakout in about 10 minutes.

## Why it matters

Willison argues the no-tools weight-reading question is more interesting than the weight-editing one. He also says a patient human could probably do the same, so the result does not show machines beyond comprehension.

## Who it's for

ML researchers and local-model users

## Try it

llm install llm-lmstudio

## Links

- [Article](https://simonwillison.net/2026/Nov/3/qwen-can-interpret-model-weights/)
- [Original Paper Blog](https://rikiyat.github.io/blog/weight-space-programming/)
- [Experiment Logs](https://github.com/RikiyaT/lm_understand-logs)