Tech News
← Home  ·  All topics

Mnist

1 GoKawiil brief on this topic · RSS for mnist

Experiment shows LLMs 'writing' neural net weights revert to pre-deep-learning tricks

A blog post describes testing 'Weight-Space Programming' by having Claude Fable 5.2 manually set all 5,994 weights of a fixed CNN for MNIST digit classification, without fitting, regression, or search. The author found the model relied on hand-crafted filters and templates reminiscent of 1990s computer vision techniques, and observed similar patterns when re-examining a Go-playing network that GPT-6 Astra built in an earlier experiment.