CURV: Coherent Uncertainty-Aware Reasoning in Vision-Language Models for X-Ray Report Generation | Read Paper on Bytez