OpenAlex · Aktualisierung stündlich · Letzte Aktualisierung: 12.03.2026, 11:11

Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.

ChatGPT Performs Worse on USMLE-Style Ethics Questions Compared to Medical Knowledge Questions

2024·9 Zitationen·Applied Clinical InformaticsOpen Access
Volltext beim Verlag öffnen

9

Zitationen

5

Autoren

2024

Jahr

Abstract

Both versions of ChatGPT performed more poorly on medical ethics questions compared to medical knowledge questions. GPT-4 significantly outperformed GPT-3.5 on overall accuracy and exhibited a significantly lower response variability in answer choices. This underscores the need for ongoing assessment of ChatGPT versions for medical education.

Ähnliche Arbeiten