Scaled results (S1 q-law, S2 deep r<=12, S3 depth, S4 brain-dump) local RTX 4000

Scaled experiments (S1 q-law, S2 deep, S3 log-depth, S4 brain-dump)

Upload README.md with huggingface_hub

Local RTX 4000 Ada results (all 5 claims): JSON + figures

initial commit