Beyond Sight: Towards Cognitive Alignment in LVLM via Enriched Visual Knowledge | Read Paper on Bytez