Table 1.

Summary table of case studies in AI applications.

Study	Models tested	Additional modifiers	Data source	Evaluation metric
1	GPT-3davinci, BARTcnn, T5, LED	None	CNN+DailyMail, booksum, sec-litigation, MIMIC-III	ROUGE, BERTscore, text length reduction
2	GPT-4, GPT-3, FLAN-T5, FLAN-UL2, Llama-2, Vicuna	ICL, QLoRA	Open-i	ROUGE, BERTscore, clinician evaluation
3	FLAN-T5, BERT, pegasus-xsum	SPeC	MIMIC-CXR	ROUGE, BERTscore
4	Med-BERT	Prediction head, fine-tuning via smaller data sets	Cerner, Truven	Disease Prediction AUC