The VARZIN Project

Mathematical Foundations

The absolute, mathematically proved core of the VARZIN program has been formalized over finite cyclic spaces.

• Affine Permutations: For N=12, exactly |Aff(Z12)| = 48 valid memory operators.
• Mirror-13 Transitions: Strongly connected transition system with 312 states and 600 directed edges.
• Convergence: Lazy random walk resolves periodicity, locking a deterministic spectral gap of γ ≈ 0.04133.

Biological & Physical Resonance

Biophysical and planetary field hypotheses failed under rigorous testing against stronger null models.

• Genomics (Track A): Chromosome 11 scan yielded 0 windows after BH-FDR under Altschul-Erickson shuffling.
• Seismicity (Track C): Clustering anomaly at Mt. Kailash dissolved (p_tectonic = 0.3956) under the Alpide Null.

Cognitive Trials

Preregistered study VPE-001A (target N=59) is planned to test whether human observers decode Core-30 axes or morpheme families above baseline.

Foundation Model Auditing (Track D)

Autoregressive LLMs (Llama 3 and Mistral) exhibit absolute structural blindness under strict semantic starvation. Clustered partitions returned ARI scores near zero (ARI_Llama = 0.0152, ARI_Mistral = -0.0337), verifying the "Morphological Hijacking" trap—they group tokens by surface spelling, not underlying geometry.

Explore the full Evidence Atlas →

Independent Computational Audit Closure (July 2026)

An independent four-stage adversarial audit has verified that the previous numeric structure-detection pipeline ("benchmark_003") did not establish a distinct, generalizable advantage beyond specialized baseline features. Consequently, all previous emergence claims built on that specific pipeline are retracted. This website serves as the corrected, evidence-first public record of the project.