LucreyaResearch › Data release

AI recommendation reliability, 2026: the data

Everything behind We audited our own AI visibility index: every raw answer, every report, every instrument. Licensed CC BY 4.0. Attribution: Vincent Wesley Couey, Lucreya.

1169 raw answer rows across 40 questions. 9 models probed for availability, 5 retired.

Files

filesize
data/reliability-report-T0-2026-09-23.json0.8 KB
data/model-mortality-2026-09-23.json2.6 KB
data/temp-probe-2026-09-23.json14.0 KB
data/size-at-temp0-2026-09-23.json16.5 KB
data/study-answers-cats-lattice-owned-T0.jsonl376.5 KB
scripts/reliability-study.mjs.txt22.7 KB
scripts/model-mortality.mjs.txt7.7 KB
scripts/temp-probe.mjs.txt6.0 KB
scripts/size-at-temp0.mjs.txt5.8 KB
scripts/build-study-page.mjs.txt31.9 KB
scripts/page-shell.mjs.txt7.7 KB
scripts/scaffold-audit.mjs.txt5.8 KB
scripts/publish-data.mjs.txt14.5 KB
bank/cats-lattice-owned.json2.9 KB
manifest.jsonn/a

Instruments are published with a .txt extension because the edge does not serve .mjs. Rename to .mjs to run them.

Stated limits

Conflict of interest. Lucreya is part of the Lattice network, which sells disclosed placements on pages that AI engines cite. That is a commercial interest in this subject. These files are published so the findings can be checked against us.
Save
Dashboard