gHashTag · gHashTag · May 12, 2026 · May 12, 2026
@@ -194,7 +194,7 @@ Trinity CLARA implements four DARPA-specified neuro-symbolic composition pattern
 
 ### References
 
-1. Coq Development - t27/codebase/proofs/ (84 theorems verifying Trinity kernel)
+1. Coq Development - t27/coq + t27/proofs (28 .v files, 218 statements = 122 Theorem + 96 Lemma, 162 Qed, 32 Admitted, 11 Abort on 2026-05-12 audit). The folkloric "84 theorems" count is retired.
 2. AR Specifications - specs/ar/*.t27 (7 specifications, 93 tests each with bounded reasoning)
 3. VSA Performance - gen/verilog/vsa/*.v (hardware synthesis files with resource utilization)
 4. ML Specifications - specs/nn/*.t27 (neural network layers, attention mechanisms, RL)

@@ -1,10 +1,20 @@
 # Current Work — Trinity t27
 
-**Last updated:** 2026-05-08
+**Last updated:** 2026-05-12
 **Note:** GF16 4×4 matmul validated on FPGA @ 323 MHz, 40350 LUTs, 64 DSP48E1, 0 latches. **TinyTapeout TTSKY26a submitted** — `gHashTag/tt-trinity-gf16`, CI running. 41.2 GOPS @ 323 MHz | 12.8 GOPS @ 100 MHz.
 
 ---
 
+## R5-PASS-4 Honest Audit (PR #594, Issue #595)
+
+- Retired the folkloric "84 theorems" claim across `research/` (actual corpus: 28 .v files, 218 statements, 162 Qed, 32 Admitted, 11 Abort — audit 2026-05-12)
+- Rewrote `docs/ZENODO.md` as a registry pointer aligned with [trios `zenodo-registry.md`](https://github.com/gHashTag/trios/blob/main/docs/infrastructure/zenodo-registry.md)
+- Bibliography honesty: commented out fake `zenodo.12345` placeholder, corrected `zenodo.19271888` mislabel (actually Koide-formula paper), and `zenodo.19377394` mislabel (actually Latin-American employment dataset) in `research/trinity-pellis-paper/G2_*`
+- Sibling PRs (R5-PASS-4): gHashTag/trinity#592 (merged), gHashTag/trinity-fpga#43 (merged)
+- Sibling PRs (R5-PASS-5): gHashTag/trinity#593, gHashTag/trinity-fpga#44 (new arXiv/ORCID class)
+- Throne: [trios#264](https://github.com/gHashTag/trios/issues/264)
+
+
 ## Active Work
 
 **GF16 Hardware Accelerator — FPGA**

@@ -1,27 +1,64 @@
 # Zenodo DOI Registry — Trinity / t27
 
-## Collection
+> **Audit 2026-05-12** — registry rewritten under R5-honest deep sweep PASS-4.
+> All March 2026 DOIs (`19224xxx`) were marked SUPERSEDED on Zenodo and replaced
+> by the May 2026 canonical set (`19227865`–`19227879`). Cite only the canonical
+> set in new work.
 
-| DOI | Title | Date | Type |
-|-----|-------|------|------|
-| [10.5281/zenodo.19224105](https://doi.org/10.5281/zenodo.19224105) | Trinity Defensive Publications — Complete Collection (69 Discoveries) | 2026-03-25 | Journal article |
-| [10.5281/zenodo.18939351](https://doi.org/10.5281/zenodo.18939351) | gHashTag/trinity — Trinity v2.0.3 FPGA Autoregressive Ternary LLM | 2026-03-11 | Software |
+## Honest framing
 
-## Individual Bundles (B001–B007)
+These DOIs are **software description stubs** on Zenodo, NOT peer-reviewed
+papers. The mathematical anchor `φ² + φ⁻² = 3` is an algebraic identity from
+`φ = (1+√5)/2`. Its Coq witness lives in this very repository
+([`gHashTag/t27/coq` + `proofs`](https://github.com/gHashTag/t27)) — **28 .v
+files, 218 statements (122 Theorem + 96 Lemma), 162 Qed, 32 Admitted, 11 Abort,
+audited 2026-05-12**.
+
+For "always-latest of B007" use **concept DOI** `10.5281/zenodo.19227876` (it
+resolves to the current B007 version regardless of revisions).
+
+## Canonical collection (curated, 2026-05-12)
+
+| DOI | Title | Type |
+|-----|-------|------|
+| [10.5281/zenodo.19227879](https://doi.org/10.5281/zenodo.19227879) | Trinity — Defensive Publications Collection (parent record) | Software description |
+| [10.5281/zenodo.18950696](https://doi.org/10.5281/zenodo.18950696) | gHashTag/trinity v2.0.3 — FPGA Autoregressive Ternary LLM | Software release |
+
+## Individual bundles B001–B007 (canonical)
 
 | DOI | Bundle | Description |
 |-----|--------|-------------|
-| [10.5281/zenodo.19224114](https://doi.org/10.5281/zenodo.19224114) | B001 | Ternary Neural Networks — HSLM 1.95M, T-JEPA, Cosine LR |
-| [10.5281/zenodo.19224115](https://doi.org/10.5281/zenodo.19224115) | B002 | Zero-DSP FPGA — Ternary MAC, DSP48E1, CORDIC, OpenXC7 |
-| [10.5281/zenodo.19224116](https://doi.org/10.5281/zenodo.19224116) | B003 | TRI-27 ISA — 36 opcodes, 27 registers, Coptic Alphabet |
-| [10.5281/zenodo.19224118](https://doi.org/10.5281/zenodo.19224118) | B004 | Queen Lotus Cycle — 6-phase self-improvement, SEVO |
-| [10.5281/zenodo.19224119](https://doi.org/10.5281/zenodo.19224119) | B005 | Tri Language — Linear Types, Algebraic Effects |
-| [10.5281/zenodo.19224120](https://doi.org/10.5281/zenodo.19224120) | B006 | Sacred Formats — GF16/TF3, phi-Distance |
-| [10.5281/zenodo.19224121](https://doi.org/10.5281/zenodo.19224121) | B007 | VSA Operations — bind/unbind/bundle, Cosine Similarity |
-
-## Specific Version DOIs
-
-| DOI | Version | Description |
-|-----|---------|-------------|
-| [10.5281/zenodo.19224139](https://doi.org/10.5281/zenodo.19224139) | v2 (latest) | Complete Collection v2 |
-| [10.5281/zenodo.18950696](https://doi.org/10.5281/zenodo.18950696) | v2.0.3 | Trinity software release |
+| [10.5281/zenodo.19227865](https://doi.org/10.5281/zenodo.19227865) | B001 | HSLM-1.95M ternary LM (training-recipe stub) |
+| [10.5281/zenodo.19227867](https://doi.org/10.5281/zenodo.19227867) | B002 | Zero-DSP FPGA — ternary inference architecture sketch |
+| [10.5281/zenodo.19227869](https://doi.org/10.5281/zenodo.19227869) | B003 | TRI-27 — ternary ISA with Coptic encoding |
+| [10.5281/zenodo.19227871](https://doi.org/10.5281/zenodo.19227871) | B004 | Queen Lotus Cycle — autonomous orchestration |
+| [10.5281/zenodo.19227873](https://doi.org/10.5281/zenodo.19227873) | B005 | Tri Language — linear types, algebraic effects |
+| [10.5281/zenodo.19227875](https://doi.org/10.5281/zenodo.19227875) | B006 | GF16/TF3 — phi-based arithmetic |
+| [10.5281/zenodo.19227877](https://doi.org/10.5281/zenodo.19227877) | B007 | VSA operations — bind/unbind/bundle |
+
+## Retired (superseded on Zenodo 2026-05-12)
+
+The following DOIs now have `[SUPERSEDED — see 10.5281/zenodo.<canonical>]`
+title prefixes and `relation: isObsoletedBy` pointing to the canonical set
+above. Do not cite them in new work.
+
+| Retired DOI | Was | Canonical replacement |
+|-------------|-----|----------------------|
+| [10.5281/zenodo.19224105](https://doi.org/10.5281/zenodo.19224105) | Complete Collection (March-2026) | `19227879` |
+| [10.5281/zenodo.19224139](https://doi.org/10.5281/zenodo.19224139) | Complete Collection v2 | `19227879` |
+| [10.5281/zenodo.19224114](https://doi.org/10.5281/zenodo.19224114) | B001 (March) | `19227865` |
+| [10.5281/zenodo.19224115](https://doi.org/10.5281/zenodo.19224115) | B002 (March) | `19227867` |
+| [10.5281/zenodo.19224116](https://doi.org/10.5281/zenodo.19224116) | B003 (March) | `19227869` |
+| [10.5281/zenodo.19224118](https://doi.org/10.5281/zenodo.19224118) | B004 (March) | `19227871` |
+| [10.5281/zenodo.19224119](https://doi.org/10.5281/zenodo.19224119) | B005 (March) | `19227873` |
+| [10.5281/zenodo.19224120](https://doi.org/10.5281/zenodo.19224120) | B006 (March) | `19227875` |
+| [10.5281/zenodo.19224121](https://doi.org/10.5281/zenodo.19224121) | B007 (March) | `19227877` |
+| [10.5281/zenodo.18939351](https://doi.org/10.5281/zenodo.18939351) | Trinity v2.0.3 FPGA LLM (older record) | use `18950696` for v2.0.3 |
+
+## Cross-sibling registry
+
+The full Trinity hive Zenodo registry (with related identifiers, communities,
+and supersession history for all 12 canonical records + 28 retired versions)
+lives in [`gHashTag/trios/docs/infrastructure/zenodo-registry.md`](https://github.com/gHashTag/trios/blob/main/docs/infrastructure/zenodo-registry.md).
+
+— end of registry —
@@ -88,15 +88,15 @@ The paper states explicitly that many relations are **empirical approximations**
 
 These Zenodo records describe **architectures and artifacts**, not theorems. Claims below should be tightened as independent benchmarks and papers appear.
 
-**Related DOIs:** [10.5281/zenodo.18939352](https://doi.org/10.5281/zenodo.18939352) (FPGA autoregressive ternary LLM), [10.5281/zenodo.19020211](https://doi.org/10.5281/zenodo.19020211) (Ouroboros), [10.5281/zenodo.19020213](https://doi.org/10.5281/zenodo.19020213) (VSA + SIMD), [10.5281/zenodo.19020215](https://doi.org/10.5281/zenodo.19020215) (phi-RoPE), [10.5281/zenodo.19020217](https://doi.org/10.5281/zenodo.19020217) (sparse ternary matmul), [10.5281/zenodo.19227877](https://doi.org/10.5281/zenodo.19227877) (VSA ops); concept [10.5281/zenodo.18947017](https://doi.org/10.5281/zenodo.18947017).
+**Related DOIs:** [10.5281/zenodo.18939352](https://doi.org/10.5281/zenodo.18939352) (FPGA autoregressive ternary LLM), [10.5281/zenodo.19020270](https://doi.org/10.5281/zenodo.19020270) (Ouroboros), [10.5281/zenodo.19020275](https://doi.org/10.5281/zenodo.19020275) (VSA + SIMD), [10.5281/zenodo.19020280](https://doi.org/10.5281/zenodo.19020280) (phi-RoPE), [10.5281/zenodo.19020282](https://doi.org/10.5281/zenodo.19020282) (sparse ternary matmul), [10.5281/zenodo.19227877](https://doi.org/10.5281/zenodo.19227877) (VSA ops); concept [10.5281/zenodo.18947017](https://doi.org/10.5281/zenodo.18947017).
 
 | ID | Claim | Domain | Status | Rationale | Artifacts |
 |----|-------|--------|--------|-----------|-----------|
 | C-ternary-001 | FPGA autoregressive ternary LLM runs inference in balanced-ternary arithmetic | HW / ML | `EMPIRICAL_FIT` | Zenodo describes design/code; independent replication + benchmarks needed. | 10.5281/zenodo.18939352 |
-| C-ternary-002 | Self-Evolving Ouroboros demonstrates a self-hosting / self-evolving cycle | Systems | `CONJECTURAL` | Need formal criteria and reproducible experiment logs. | 10.5281/zenodo.19020211 |
-| C-ternary-003 | VSA balanced ternary + SIMD gives stable high-dimensional VSA ops | VSA / numerics | `EMPIRICAL_FIT` | Zenodo description; needs stability tests vs binary VSA baselines. | 10.5281/zenodo.19020213, 10.5281/zenodo.19227877 |
-| C-ternary-004 | phi-RoPE improves quality/stability vs standard RoPE on binary models | ML / attention | `CONJECTURAL` | Need public perplexity / stability / spectral comparisons. | 10.5281/zenodo.19020215 |
-| C-ternary-005 | Sparse ternary matmul wins FLOPs/W and/or latency on FPGA vs dense binary matmul | HW | `CONJECTURAL` | Need published measurement methodology. | 10.5281/zenodo.19020217 |
+| C-ternary-002 | Self-Evolving Ouroboros demonstrates a self-hosting / self-evolving cycle | Systems | `CONJECTURAL` | Need formal criteria and reproducible experiment logs. | 10.5281/zenodo.19020270 |
+| C-ternary-003 | VSA balanced ternary + SIMD gives stable high-dimensional VSA ops | VSA / numerics | `EMPIRICAL_FIT` | Zenodo description; needs stability tests vs binary VSA baselines. | 10.5281/zenodo.19020275, 10.5281/zenodo.19227877 |
+| C-ternary-004 | phi-RoPE improves quality/stability vs standard RoPE on binary models | ML / attention | `CONJECTURAL` | Need public perplexity / stability / spectral comparisons. | 10.5281/zenodo.19020280 |
+| C-ternary-005 | Sparse ternary matmul wins FLOPs/W and/or latency on FPGA vs dense binary matmul | HW | `CONJECTURAL` | Need published measurement methodology. | 10.5281/zenodo.19020282 |
 
 ---
 
@@ -105,7 +105,7 @@ These Zenodo records describe **architectures and artifacts**, not theorems. Cla
 | ID | Claim | Domain | Status | Rationale | Artifacts |
 |----|-------|--------|--------|-----------|-----------|
 | C-meta-001 | Trinity / t27 is a spec-first ternary stack; Zig/C/Verilog backends are generated from `.t27` | PL / compilers | `EMPIRICAL_FIT` | Repo layout + CI (gen headers, conformance) demonstrate discipline; full `docs/nona-02-organism/LANGUAGE_SPEC.md` + backend contracts still incomplete. | This repo; `docs/nona-02-organism/LANGUAGE_SPEC.md`, `docs/BACKEND_CONTRACT.md`. |
-| C-meta-002 | Trinity / t27 is self-hosting / self-evolving | Systems | `CONJECTURAL` | Define terms precisely + reproducible pipeline; partial story in rings + Ouroboros Zenodo. | 10.5281/zenodo.19020211; `CANON.md`, `docs/nona-01-foundation/SEED-RINGS.md`. |
+| C-meta-002 | Trinity / t27 is self-hosting / self-evolving | Systems | `CONJECTURAL` | Define terms precisely + reproducible pipeline; partial story in rings + Ouroboros Zenodo. | 10.5281/zenodo.19020270; `CANON.md`, `docs/nona-01-foundation/SEED-RINGS.md`. |
 
 ---
 
@@ -134,10 +134,10 @@ These Zenodo records describe **architectures and artifacts**, not theorems. Cla
 | [10.5281/zenodo.18947017](https://doi.org/10.5281/zenodo.18947017) | Concept DOI (all versions) | 2026-03-10 |
 | [10.5281/zenodo.18950696](https://doi.org/10.5281/zenodo.18950696) | Latest Trinity Framework version | 2026-03-10 |
 | [10.5281/zenodo.18939352](https://doi.org/10.5281/zenodo.18939352) | FPGA Autoregressive Ternary LLM | 2026-03-10 |
-| [10.5281/zenodo.19020211](https://doi.org/10.5281/zenodo.19020211) | Self-Evolving Ouroboros | 2026-03-14 |
-| [10.5281/zenodo.19020213](https://doi.org/10.5281/zenodo.19020213) | VSA Balanced Ternary + SIMD | 2026-03-14 |
-| [10.5281/zenodo.19020215](https://doi.org/10.5281/zenodo.19020215) | phi-RoPE Attention | 2026-03-14 |
-| [10.5281/zenodo.19020217](https://doi.org/10.5281/zenodo.19020217) | Sparse Ternary MatMul | 2026-03-14 |
+| [10.5281/zenodo.19020270](https://doi.org/10.5281/zenodo.19020270) | Self-Evolving Ouroboros | 2026-03-14 |
+| [10.5281/zenodo.19020275](https://doi.org/10.5281/zenodo.19020275) | VSA Balanced Ternary + SIMD | 2026-03-14 |
+| [10.5281/zenodo.19020280](https://doi.org/10.5281/zenodo.19020280) | phi-RoPE Attention | 2026-03-14 |
+| [10.5281/zenodo.19020282](https://doi.org/10.5281/zenodo.19020282) | Sparse Ternary MatMul | 2026-03-14 |
 | [10.5281/zenodo.19227877](https://doi.org/10.5281/zenodo.19227877) | VSA Operations for Ternary Computing | — |
 
 ---

@@ -374,7 +374,7 @@ This work emerged from discussions within the Trinity S³AI research group. We a
 
 **[chimera2026]** S. Pellis, *CKM Wolfenstein Parameters via Golden Ratio Polynomials*, preprint, 2026.
 
-**[olsen2026]** S. Olsen, *Historical Context of φ in Physics: from Pythagoras to Bohm*, Zenodo, DOI: 10.5281/zenodo.19377394, 2026.
+**[olsen2026]** S. Olsen, *Historical Context of φ in Physics: from Pythagoras to Bohm*, contribution to this paper (2026). [R5-honest audit 2026-05-12: the DOI 10.5281/zenodo.19377394 previously cited here resolves to an unrelated Latin-American employment dataset, not to Olsen's work. Restated as personal contribution; replace with real Olsen Zenodo DOI before publication.]
 
 **[PDG2024]** Particle Data Group (S. Navas et al.), *Review of Particle Physics*, Phys. Rev. D **110**, 030001 (2024).
 

@@ -513,8 +513,11 @@ \section*{Acknowledgments}
 \bibitem{olsen2026}
 S.~Olsen,
 \textit{Historical Context of $\varphi$ in Physics: from Pythagoras to Bohm},
-\textit{Zenodo},
-\href{https://doi.org/10.5281/zenodo.19377394}{DOI:~10.5281/zenodo.19377394}, 2026.
+% R5-honest audit 2026-05-12: the DOI 10.5281/zenodo.19377394 previously
+% cited here does NOT resolve to an Olsen paper on phi. It resolves to an
+% unrelated Latin-American employment dataset. Aligned with V0.9 of this
+% paper, the citation is restated as a personal contribution.
+contribution to this paper (2026).
 
 \bibitem{PDG2024}
 Particle Data Group (S.~Navas et al.),

@@ -680,8 +680,13 @@ \section*{Acknowledgments}
 \bibitem{olsen2026}
 S.~Olsen,
 \textit{Historical Context of $\varphi$ in Physics: from Pythagoras to Bohm},
-Zenodo,
-\href{https://doi.org/10.5281/zenodo.19377394}{DOI:~10.5281/zenodo.19377394}, 2026.
+% R5-honest audit 2026-05-12: the DOI 10.5281/zenodo.19377394 previously
+% cited here does NOT resolve to an Olsen paper on phi. It resolves to
+% ``Visual Summary (Français): Empleo en América latina: Base de Données
+% par Pays (2000-2020)'' — an unrelated Latin-American employment dataset.
+% Aligned with V0.9 of this paper, the citation is restated as a personal
+% contribution; replace with a real Olsen Zenodo DOI before publication.
+contribution to this paper (2026).
 
 \bibitem{PDG2024}
 Particle Data Group (S.~Navas et al.),

@@ -31,7 +31,11 @@
         $^3$ Wisdom Traditions Center, LLC, 1802 SW 28th St., Ocala, Florida 34471 USA}\\[2pt]
 {\small \texttt{admin@t27.ai} \quad \texttt{sterpellis@gmail.com}}
 \date{April 2026}
-\doi{https://doi.org/10.5281/zenodo.12345}
+% DOI placeholder removed by R5-honest audit 2026-05-12. The literal
+% `10.5281/zenodo.12345` resolved to an unrelated paper on Roman history
+% (Clientelas, relaciones internacionales e imperialismo en la expansión
+% de la República romana). Replace with a real Zenodo DOI before publication.
+% \doi{https://doi.org/10.5281/zenodo.XXXXXXXX}
 
 \begin{document}
 \maketitle
@@ -1148,8 +1152,14 @@ \section*{Acknowledgments}
 PhilArchive (2026).
 
 \bibitem{zenodo19271888}
-Zenodo inverse participation ratio reference,
-Zenodo DOI:~10.5281/zenodo.19271888 (2026).
+% R5-honest audit 2026-05-12: DOI 10.5281/zenodo.19271888 resolves to
+% McKoy, C. (2026) ``V2: The Cogito of Fermion Masses: Two Independent
+% Mathematical Framings of the Koide Formula ... and Its Identification
+% with the AdS/CFT Correspondence''. The original `inverse participation
+% ratio reference' label was incorrect; the cited record is about Koide.
+McKoy, C.
+\textit{V2: The Cogito of Fermion Masses: Two Independent Mathematical Framings of the Koide Formula, a New Geometric Interpretation, and Its Identification with the AdS/CFT Correspondence},
+Zenodo DOI:~\href{https://doi.org/10.5281/zenodo.19271888}{10.5281/zenodo.19271888} (2026).
 
 \bibitem{kagome2026}
 Kagome lattice UCD result (2026).

@@ -89,7 +89,7 @@
 spanning quantum gravity, neural dynamics, and cosmology.
 The central named constant is $\aph = \ph^{-3}/2 = (\sqrt{5}-2)/2\approx0.118034$,
 coinciding with $\alpha_s(m_Z)=0.1180\pm0.0009$ within $0.03\sigma$.
-Machine-verified Coq proof base: \textbf{84 theorems}, 13 compiled \texttt{.v}~files
+Machine-verified Coq proof base: \textbf{audited 2026-05-12: 28 \texttt{.v} files, 218 statements (122 Theorem + 96 Lemma), 162 Qed, 32 Admitted, 11 Abort}. The earlier folkloric ``84 theorems'' count is superseded by this audit
 (Rocq~9.1.1 + \texttt{coq-interval}$\ge$4.8.0, 0~compilation errors).%
 \footnote{Repository: \url{https://github.com/gHashTag/t27}.
 Monte Carlo significance: $p=1.47\times10^{-53}$ (Poisson, $\mu_0\approx0.4$).
@@ -484,7 +484,7 @@ \section*{5.\quad Chimera v1.0: ML Search + Coq Verification}
 \end{tikzpicture}
 \caption{L1--L7 derivation hierarchy. Each level adds mathematical structure;
 formulas at higher levels derive from lower levels via composition.
-Total: 84 theorems across 7 levels.}
+Total: audited 2026-05-12 as 218 statements (162 Qed, 32 Admitted, 11 Abort); the earlier ``84 theorems / 7 levels'' shorthand is retired.}
 \label{fig:levels}
 \end{figure}
 
@@ -764,7 +764,7 @@ \section*{12.\quad Machine-Verified Proof Base}
 The ML search engine (Python, 2400+ lines) generates Trinity monomial
 candidates, which are then certified in Coq. This AR+ML composition
 demonstrates the CLARA technical objective of verifiable reasoning
-at scale: 84 theorems, L1-L7 hierarchy, depth $\le$ 7 levels.
+at scale: 218 audited statements (2026-05-12), L1-L7 hierarchy, depth $\le$ 7 levels.
 
 % ============================================================
 \section*{13.\quad Conclusion}