Трофімов O. and H. Hulak, “Method for enhancing the robustness of large language models against adversarial attacks”, Bezpeka ìnf., vol. 32, no. 2, pp. 72–82, Aug. 2026.