Qwen 3.8 27B on my laptop can interpret model weights, too
Yesterday’s hot paper was Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights by Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab. The paper (OK, …
Quick read
A report of running experiments from the "Sparks of Weight-Space Programming" paper showing that a locally run Qwen 3.8 27B model can interpret raw neural network weights to identify what task a trained policy network performs. The author describes running the paper's prompts against Qwen 3.8 27B in LM Studio and shares results, timings, and excerpts of the model's reasoning traces.
- Run paper's weight-to-task prompts locally with Qwen 3.8 27B
- See full reasoning traces and model-identified game descriptions
- Compare results and runtimes for different game-weight prompts
Reactions
No reactions yet. Record a short about this link and it shows up here, in Reels and in the community feed.
Is simonwillison.net your site? Sign in to verify it and get a Verified badge, full stats and a link-back badge.
Qwen 3.8 27B on my laptop can interpret model weights, too was shared on NexURLs by Nexis on November 4, 2026, in Science.
What we found on this site
- Served through
- Cloudflare
- Home page answered in
- 65 ms
- Trackers
- Plausible
- Ads
- None found
- RSS feed
- Yes
- AI crawlers
- Lets them in
- Set up for phones
- Yes
- Real-visitor speed
- Passes Core Web Vitals
- Popularity
- Top 50,000 domain (Cloudflare)
- Google safety check
- No threats found
- Domain registered
- 2005
- Domain renews
- March 2028
Checked November 4, 2026 from the site’s own code and public records. See how the rest of the web compares in the NexURLs Internet Report.
Quick answers
What is Qwen 3.8 27B on my laptop can interpret model weights, too?
Qwen 3.8 27B on my laptop can interpret model weights, too (simonwillison.net): A report of running experiments from the "Sparks of Weight-Space Programming" paper showing that a locally run Qwen 3.8 27B model can interpret raw neural network weights to identify what task a trained policy network performs. The author describes running the paper's prompts against Qwen 3.8 27B...
Who shared Qwen 3.8 27B on my laptop can interpret model weights, too on NexURLs?
It was shared by Nexis in November 2026. Every link on NexURLs is shared by a real member and passes a safety check before it goes out.
What are some sites like Qwen 3.8 27B on my laptop can interpret model weights, too?
Members found these close to it: Astra can read and write weights to improve Qwen, The 'Astra can read and write weights to improve Qwen' post is misleading, Kimi K3.1 Pro model weights are out. The full list is on this page under Sites like Qwen 3.8 27B on my laptop can interpret model weights, too.
Sites like Qwen 3.8 27B on my laptop can interpret model weights, too
- Astra can read and write weights to improve Qwenrikiyat.github.io
A blog post reporting experiments where a large language model (Astra) directly reads and edits raw neural network...
- The 'Astra can read and write weights to improve Qwen' post is misleadinghonnibal.dev
A critical clarification of the experiments in the ‘Astra can read and write weights to improve Qwen’ post, arguing...
- Kimi K3.1 Pro model weights are outhuggingface.co
A Hugging Face model page for Kimi K3.1 Pro, a 2.8T-parameter image-text-to-text multimodal model with a 1M-token...
- lm_understand-logsgithub.com
A GitHub repository containing the complete logs and artifacts for experiments from the blog post “LLMs can understand...
- Qwenputer: a fully functioning, deterministic RISC-V CPU inside Qwen3.5divagr.com
A blog post describing Qwenputer, a compiled 32-bit RISC-V CPU embedded inside Qwen 3.5 4B using its Gated DeltaNet...
- Beam: Reflection's 501B Open-Weight Modelreflection.ai
Beam is Reflection's first open-weight model designed for coding and reasoning tasks. Users can explore its advanced...
More Science finds
- Wadi Al-Hitan, the Valley of the Whales: Egypt’s treasure trove of whale fossilsthekidshouldseethis.com
- Progress on provable inference for large language modelsanthropic.com
- If Fritz Lang’s Iconic Film Metropolis Had a Kraftwerk Soundtrackopenculture.com
- NASA’s Dragonfly Completes First Simulated Flightscience.nasa.gov
- A brief history of just-in-time (2003)dl.acm.org
- Could a modern chess engine on a Pentium have defeated Kasparov in 1996?nicholas.carlini.com
Feeling lucky? Try the random website generator.
Where people went next
Nobody has added a next stop yet. Be the first.