CM MODEL CORRECTNESS AUDIT — UNPUBLISHED SUPPORTING PACKAGE, 12 NOVEMBER 2026
Start with the model-boundary summary and correction ledger, then the unchanged 28-page September 18 report.
Placement: supporting material beneath Correspondence Matrices / CM-LM; not a replacement foundations manuscript or a new top-level research family.
The original byline is Prepared for Brian / Quantum CM project. No new author, reuse license, external review, hardware validation or novelty clearance is inferred.
source.zip contains the original standalone source and modular source/table closure. Modular expansion agrees with standalone after whitespace normalization. No new TeX compilation is claimed.
The source-and-evidence archive preserves the original audit package, excluding interpreter caches. Historical source snapshots are superseded audit inputs, not current recommendations. Start with this README, not old source-snapshot summaries.
The old root manifest has eight differences explained by CRLF/LF and one unresolved timing-file mismatch. All three nested historical manifests match exactly. Original manifests and bytes are preserved; the new release manifest records current packaging identities. Timing logs are not performance evidence.
The original report records 59 historical and 31 independent tests, plus large driver campaigns. New preparation checks are recorded separately in verification.json; they do not recertify all inherited computations. Do not run run_all.py merely to inspect the package: it regenerates evidence.
The full reader is the original PDF. HTML supplies a new reading guide and correction ledger, not a transcription of all equations. Extracted text is approximate.
Nothing in this package establishes physical quantum computation, a cognitive architecture, consciousness, compression or speedup.
