three-tier-architecture-completeness

OUT derived (depth 8)

Created 2026-08-25T03:13:05+00:00

The three-tier memory architecture (broad superposition read, narrow parametric write, unbounded context write) constitutes a complete and faithful model of LLM knowledge management

Justifications

SL — three-tier-memory-architecture and context-as-infinite-dimension-complement jointly assert the architecture is complete. However, ROME-edited GPT-J answering only 7.4% of multi-hop questions (down from 40.5%) reveals a compositional-knowledge gap the three tiers do not cover. The architecture claim holds only while this multi-hop failure is OUT (i.e., resolved or shown to be irrelevant to the single-fact scope).

Antecedents (all must be IN):

  • OUT three-tier-memory-architecture — The LLM implements a three-tier memory architecture: broad superposition-based read, narrow rank-one parametric write confined to the covariance-whitened addressable subspace, and unbounded contextual write via the context window—with the geometric boundary of C⁻¹k* defining the precise demarcation between parametric and contextual regimes.
  • OUT context-as-infinite-dimension-complement — The context window is the operational realization of the "infinite" or "unbounded" dimension in the superposition framework: while the d-dimensional residual stream is over-compressed to store N≫d features (necessitating covariance whitening for well-defined read/write operations), the context window provides an unbounded, interference-free channel where knowledge is injected without superposition, without the need for C⁻¹k* addressability, and without the rank-one constraint—making it the architectural complement to the finite, superposed parametric memory.

Unless (any of these IN defeats this justification):