Recommendation Diversity is an important evaluation metric for recommendation diversity and novelty tasks in the field of natural language processing. Diversity measures the degree of difference between items in a recommendation list and reflects the richness of recommendation results in terms of content or features. It is typically reported as a dimensionless score. This metric is widely used in relevant benchmarks and papers and can effectively reflect model performance on this dimension. Higher values are better. It helps researchers and engineers quantify differences in actual model performance in recommendation diversity and novelty scenarios, facilitating horizontal comparison of different methods or model versions.