context-extends-semantic-currency-to-infinite-dimension
OUT derived (depth 7)
Created 2026-08-25T03:45:56+00:00 · Reviewed 2026-08-25T04:02:18+00:00
The context window extends the "universal semantic currency" (covariance geometry) into an effectively unbounded-dimensional space, making the full LLM read/write system a finite-dimensional-plus-infinite-dimensional geometric object rather than a purely finite one.
Justifications
SL — The geometry-as-currency claim defines the finite d-dimensional space; the unbounded-complement claim establishes context as the infinite extension. Together they show the full semantic currency is not merely d-dimensional but d+∞, with the finite part being the editable/measurable core and the infinite part being the contextual write channel. Neither claim alone gives the full geometric picture.
Antecedents (all must be IN):
- OUT geometry-as-universal-semantic-currency — The covariance geometry is the single operational definition of "meaning" in LLMs, simultaneously determining what can be measured (evaluation via cosine/Spearman), what can be modified (rank-one editing via C⁻¹k*), what converges across architectures (SAE/Park universality), and what is hierarchically structured (feature neighborhoods as theorem instantiations)—making it a model-independent semantic currency rather than an architecture-specific artifact.
- OUT context-write-unbounded-complement — The context window functions as the unbounded write channel that complements the rank-one parametric write: where parametric editing is confined to a single MLP's addressable subspace, the context channel injects arbitrary quantities of verified knowledge, making the two write channels capacity-complementary rather than redundant.
Dependents
These beliefs depend on this one:
- OUT context-completes-tail-coordinate-space — The context window is not "more memory" but the geometric completion of the addressable coordinate system: it specifically fills the tail *directions* absent from the rank-one parametric subspace, making the full semantic space (parametric + contextual) a closed coordinate system rather than two disjoint mechanisms.