Growing Intelligence
Research

published, archived,
and citable

Every claim on this site traces to an archived record with a DOI. Where we have no archived result, we publish no number.

Archived records
Preprint2026-04-28

Holosynthics: Composition Laws for Signed Digital Entities

The composition theory the whole product line rests on — the laws that govern how signed digital entities combine.

DOI 10.5281/zenodo.19852341

Preprint2026-04-30

Verifiable Process Infrastructure: A Formal Framework for Cryptographically Provable Business Processes

Formal definition of Verifiable Process Infrastructure: four axioms, three theorems establishing soundness, necessity and sufficiency, and eleven regulatory regimes — including SOX, DORA, SEC 17a-4, MiFID II, GDPR and BCBS 239 — shown to require properties no existing tool provides. The GI Engine is its reference implementation.

DOI 10.5281/zenodo.19921021

Preprint2026-06-17

Empirical Validation of the Holosynthics Periodic Table: Cross-Model Universality, Hallucination Detection, and Surgical Repair Across Transformer Architectures

From the archived preprint (DOI 10.5281/zenodo.20736159): The paper reports 3,000 sequential edits, ×9.08 surgical repair on Llama 3.1 8B, and validation across 11 transformer models from 9 organizations.

DOI 10.5281/zenodo.20736159

Report2026-07-16

BIO56: Deterministic Decomposition of Biological Data into 56 Elements — Experimental Results

941 biological entities across 4 independent data domains decomposed into 54,986,846 atoms by a deterministic 7-law classification cascade.

DOI 10.5281/zenodo.21394366

Report2026-07-17

ROBO56: Universal Robotic Sensor Decomposition — 10 Sensors, 2 Platforms, 74.4M Atoms

Robotic sensor data from an autonomous vehicle and a drone decomposed into 74,373,570 atoms by the same engine proven on biological data.

DOI 10.5281/zenodo.21409175

Case studies

three production models, certified

32B general model32% → 70%0 collateral
24B banking model32% → 70%0 collateral
24B healthcare modelcertified0 collateral

Every scan runs against a world-standard benchmark suite spanning 49 categories — covering knowledge, reasoning, math, logic, reading, hallucination, safety, and commonsense, plus domain-specific categories. Full case studies available on request.

Provenance

signed, and checkable by anyone

Every result referenced on this site is signed by the GI Engine, so an auditor can verify that a finding came from us and was not altered after the fact. Classification, detection, and repair methods. Protected by patents filed with USPTO. Application numbers are available on request under NDA.

Research FAQ

questions

Where is the underlying evidence published?
Openly on Zenodo. Five records are archived: the Holosynthics composition-laws preprint, the Verifiable Process Infrastructure framework, the cross-model empirical validation, and the BIO56 and ROBO56 experimental reports. Each is a citable, versioned record rather than a marketing summary.
What exactly counts as collateral damage?
A previously-correct answer that becomes incorrect after a repair. It is measured on a held-out control set of items the model answered correctly before the edit, scored under a no-teacher-forcing protocol so the model is not cued toward the gold answer during evaluation.
How does Probe56 repair differ from knowledge editing?
Knowledge editing injects a new fact by modifying weights, and the published literature reports that those edits propagate into unrelated facts. Probe56 performs targeted parameter-level repair instead: each edit is verified against the full control set and ships with a measured collateral report, and any edit that degrades a previously-correct answer is rejected automatically. Zero measured collateral, verified and signed.
Does the same engine work outside language models?
Yes, and that is the strongest evidence that the decomposition is universal rather than tuned. The same 8-atom-kind engine produced 54,986,846 atoms from biological data across 4 domains and 74,373,570 atoms from robotic sensor data across 2 platforms sharing no hardware.