2026. Adaptive Defense Framework for Mitigating Prompt Injection Attacks in Large Language Models. International journal of applied mathematics. 6, 07 (Jul. 2026), 01–10.