Data Pack
Local ML warmup outputs.
Minimal data pack for small warmup experiments. This is not a benchmark, just a repeatable local check of structural limits.
Related writing: Scale Raises Confidence, Not Capacity.
Live on Hedegreen Research
ML Warmups Data
Perceptron gate results and a minimal XOR two-layer run, published as direct JSON outputs.
Two rooms only. Enough to prove the core claim: scaling increases confidence in what a structure can already do, not what it cannot.
If you need scripts or method context, see the research pack under editorial/rooms/ml-warmups/ in the repo.