ResearchBenchmarksLexical Diversity on PubMedFollow72Self-BLEUAUG-PE16.8831.1945.559.81Apr 8, 2026Evaluation ResultsMethodMethodLinksSelf-BLEUN-Gram ScoreAUG-PEModel=GPT-4o-miniModel=GPT-4o-mini2026.047228AUG-PEModel=GPT-3.5Model=GPT-3.52026.046234AUG-PEModel=Phi-4Model=Phi-42026.045139RPSGModel=GPT-4o-miniModel=GPT-4o-mini2026.043353RPSGModel=GPT-3.5Model=GPT-3.52026.042957RPSGModel=Phi-4Model=Phi-42026.042763RPSGModel=DeepSeek-R1Model=DeepSeek-R12026.042081AUG-PEModel=DeepSeek-R1Model=DeepSeek-R12026.041974