sparticle62

That's the closest prior work, yes, and the memory half sits squarely in that lineage. Three differences worth naming. Reads are content-addressed in the Poincare ball rather than by cosine similarity in R^n, which is what lets a bounded top-4 read cover a hierarchy of contexts instead of a flat neighborhood. There is no controller emitting explicit read and write heads; writes are novelty-gated with a refractory counter that rate-limits overwriting the same slot, so repeated contradictory updates do less damage. And the fast weights do not stay external forever, they get consolidated into the recurrent transition matrix by closed-form ridge regression.

The backbone is also a selective SSM rather than an LSTM controller, so the sequence half is closer to Mamba than to an NTM. DNC comparisons are fair too, and I should have cited both in the readme.