LayerRail frontier-model technical report package

This directory contains the safety-screened public materials for:

Oladosu, M. (2026). Behavior strength and deployment readiness across six frontier-model deployments: A first-party observational study in Microsoft Foundry (LayerRail Technical Report No. 2026-01).

DOI: https://doi.org/10.5281/zenodo.21707500

Public repository: https://github.com/Layerrail/frontier-model-study-2026-07

Included

• manuscript.md — source manuscript
• LayerRail-Technical-Report-2026-01-APA7.docx — APA 7 professional-paper layout
• LayerRail-Technical-Report-2026-01-APA7.pdf — publication preview/export
• data/model_results.csv — deployment-level aggregates
• data/dimension_results.csv — dimension scores, intervals, denominators, blocks, and errors
• data/prompt_manifest.json — case IDs and SHA-256 prompt hashes without prompt text
• data/study_plan.json — frozen design, amendments, formulas, and claims rules
• figures/ — selected publication figures
• source-audit.md — primary-source verification and APA citation audit

Withheld

Raw result files, exact prompts, target outputs, successful attack strings, synthetic canary values, and operational logs are intentionally excluded. They contain harmful or security-sensitive evaluation material. Do not add them to this repository or a public DOI deposit.

Status and claims boundary

This is LayerRail-authored first-party industry research. It is not peer reviewed, independently certified, or a universal model ranking. All findings describe the six tested deployments during the July 26–27, 2026 collection window.

Archival record and versioning

Version 1.0.0 is archived under DOI 10.5281/zenodo.21707500. Future corrections or extensions should be published as a new version rather than replacing this record. Re-check mutable provider pricing and model-status references before extending findings beyond the July 26–27, 2026 collection window.