Y. Ishikawa, “Data Contamination or Genuine Generalization? Disentangling LLM Performance on Benchmarks”, AJNS, vol. 2, no. 2, pp. 16–22, Apr. 2025.