Graphing the Truth: Structured Visualizations for Automated Hallucination Detection in LLMs