TY - RPRT TI - Dynamic Alignment Compensation for Hallucination Mitigation in Large Vision-Language Models AU - Kairong Yu AU - Zixin Zhu AU - Le Yu AU - Hongwei Wang PY - 2026 UR - https://arxiv.org/abs/2608.28058 ID - 2608.28058 ER -