research-progressionThe research, as it developedThe archive contains mathematical proposals, several software lineages, original telemetry, a later declaration manuscript, and a completed partition-search experiment. Their relationship is part of the research record.Research methodsResearch review◇↗
normalized-mip-studyExact partitions and measured search performanceA completed benchmark on 60 synthetic recurrent networks, with 163,140 reference bipartitions and explicit classical and simulated-quantum budgets.Research methodsResearch review◇↗
declaration-studyDeclaration and persistent state changeThe latest recovered manuscript distinguishes a proposed declaration mechanism from what one public adapter-training trajectory actually measures.Research methodsResearch review◇↗
fibonacci-manuscript-reviewThe Fibonacci manuscript and its mathematical reviewThe original equations remain readable. The review explains which mathematical properties hold, which interpretation is unvalidated, and why the later manuscript changed its approach.Research methodsResearch review◇↗
g16-mathematicsReading G16, DNA and the implemented mathematicsThe names organize the project. The equation pages explain the actual operations, their domains, their version differences and the claims those operations can support.Research methodsResearch review◇↗
evidence-methodHow to read the evidenceSource recovery, mathematical explanation, saved-result verification, runtime tests and scientific interpretation answer different questions.Research methodsResearch review◇↗
donor-model-dossiersDonor models, local arithmetic, and the August source dossiersKimi K2.5 and DeepSeek-R1 connect the project to pretrained model architectures. Their local dossiers preserve exact code and reported inventories while revealing several concrete summary-versus-source differences.Research methodsResearch review◇↗