Echoes in AI: Quantifying lack of plot diversity in LLM outputs

scores among story segments correlates moderately with human judgment of surprise level, even though score computation is completely automatic without relying on human judgment.