Replication and stress tests of Sparks of Weight Space Programming: LLMs can understand and write neural network weights (post, logs). Blog post: I Tried to Break Weight Programming, by Xuanqiang ...
GPT-6 Astra edited Qwen’s weights and raised MATH-500 from 7.9% to 48.1% in author-run tests. How weight-space programming works—and its limits.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results