Ishikawa, Y. Data Contamination or Genuine Generalization? Disentangling LLM Performance on Benchmarks. AJNS 2025, 2, 16-22.