Reproducibility Package: Surrogate Testing of Normalized Prime Gaps via MI/CMI under IAAFT Nulls

This archive contains the full reproducibility package for the manuscript: Information-Theoretic Surrogate Testing of Normalized Prime Gaps: Mutual and Conditional Mutual Information Under IAAFT Nulls (Angelina Hintsanen, NEXUS AI Lab). The package includes Python scripts, configuration metadata, CSV outputs, and publication-quality figures for three experiment phases: Phase B1: Mutual information (MI) lag sweep on normalized prime gaps with IAAFT surrogate hypothesis testing Phase B2: MI surrogate testing on four control sequences (null validation suite) Phase B3: Conditional mutual information (CMI) lag sweep with growing-conditioning structure test All outputs are provided as produced by the pipeline: per-lag result tables, run metadata JSON, and PNG/PDF plots. The archive is designed for independent replication and extension. Contents PrimeGap_PhaseB1_MI_* (results + metadata + figures) PrimeGap_PhaseB2_controls_* (controls + summary tables + figures) PrimeGap_PhaseB3_CMI_* (results + metadata + figures) prime gaps, primes, information theory, mutual information, conditional mutual information, surrogate data, IAAFT, nonlinear dependence, computational number theory, experimental mathematics, kNN estimator, KSG estimator, Frenzel-Pompe