Answers are grounded in this paper and cite their sources.
We give explicit covering arrays that improve the best known size for 31 parameter sets: 13 of strength 3 and 18 of strength 4. The note describes the constructions, the verification by two separately written checkers, the use of a language model as a tool, a baseline without it and a pre-registered prediction that failed. Version 1.1 corrects version 1.0, which claimed 39 parameter sets: eight had already been overtaken by a deposit of Bai XunDi (doi:10.5281/zenodo.23787055), credited in the note.
No takes yet. Share an insight, caveat, or question.
Theo Nicolas Sitjar (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: