// RESEARCH LIBRARY

Research Library.

Whitepapers, canonical research papers, engagement papers, technical notes, briefings, presentations, and case studies from the SORT and SORT-AI research program. The library is organized as a research hierarchy, from core framework foundations to domain architecture, application papers, operational diagnostics, and executive-facing materials.

Whitepapers 6 editions

Core SORT whitepaper line in two publication layers: canonical reference editions preserved in the Zenodo archive, and journal-style preprint editions for formal scientific distribution. Versions are cumulative — each builds on and extends the preceding architecture.

Canonical Reference Whitepapers

Journal Preprint Whitepapers

Research Papers 13 papers

Canonical SORT-AI research line. The primary domain architecture paper anchors the program, followed by the Core-3 application papers, infrastructure and runtime research, and safety, alignment, and structural risk papers. Published via Preprints.org, Zenodo, and SSRN.

Domain Architecture & Strategic Framing

Core-3 Application Papers

Infrastructure, Runtime & Efficiency

Safety, Alignment & Structural Risk

PREPRINT · SSRN

SORT-AI: A Projection-Based Structural Framework for AI Safety, Alignment Stability, Drift Detection, and Scalable Oversight

Projection-based safety module built on a closed algebra of 22 idempotent operators. Provides diagnostics for drift accumulation, operator collapse, invariant violation, and destabilization of alignment-relevant fixed points.

Open DOI →
PREPRINT · PREPRINTS.ORG

SORT-AI: A Structural Safety and Reliability Framework for Advanced AI Systems with Retrieval-Augmented Generation as a Diagnostic Testbed

Structural safety framework modeling AI systems as operator chains under global consistency constraints. Uses retrieval-augmented generation as a testbed for analyzing hallucination, mis-grounding, and deceptive stability.

Open DOI →
PREPRINT · SSRN

An Operator-Projection Framework for Structural Risk Assessment in Advanced AI Systems

Mathematically hardened operator framework for analyzing stability, drift, fixed-point structure, and emergent non-local interactions in large AI models. Supports analysis of mesa-optimization conditions and misalignment trajectories.

Open DOI →
ZENODO ARCHIVE

Structural Oversight as a Performance Advantage: Why Frontier AI Governance Should Optimize for Governable Capability

Engagement paper on frontier AI governance as a structural diagnostics problem. Treats governance itself as the use case — distinguishing blunt capability suppression from oversight architectures that preserve useful capability by making deployed behavior auditable, controllable, and productive.

Open DOI →

Engagement Papers, Runtime Notes & Case Studies 8 documents

Applied SORT-AI engagement papers, runtime notes, operational diagnostics, and case studies. These documents translate the research framework into concrete architectural readings of runtime instability, benchmark divergence, recursive agentic execution, semantic failure, and control-surface risk. All entries are archived on Zenodo.

Runtime & Operational Technical Notes

Case Studies

Presentations 16 documents

Finalized presentation decks across four working layers: framework foundations and core applications, runtime and execution geometry, agentic systems and control surfaces, and governance, economy, and strategic stability. Designed for AI engineers, infrastructure architects, hyperscaler runtime teams, and executive decision audiences.

Framework & Core Applications

Runtime, Inference & Execution Geometry

Agentic Systems & Control Surfaces

Governance, Economy & Strategic Stability