ChatGPT-4.0 Achieves 85% in Clinical Neurology Exam.
OpenAI's ChatGPT-4.0 has achieved a groundbreaking milestone by answering 85% of questions from the American Board of Psychiatry and Neurology correctly. This success implies significant potential for the technology in the field of clinical neurology.
Key Points:
The latest update, ChatGPT-4.0, underwent a clinical neurology exam, showcasing an impressive 85% accuracy in a proof-of-concept study.
The research, conducted by a team from the University Hospital Heidelberg and the German Cancer Research Center, was published on Dec. 7, with the test performed on May 31.
Comparatively, the older version, ChatGPT-3.5, scored 66.8%, while ChatGPT-4.0 excelled with 85%, outperforming human users in various neurology-related questions.
The study suggests potential applications in clinical neurology after further refinements to the large language models (LLMs).
Researchers emphasize the need for caution in practical usage due to imperfections in high-order cognitive tasks, indicating ongoing development and fine-tuning for optimal application in clinical neurology.
This achievement serves as a proof-of-concept for LLM capabilities, paving the way for potential advancements in healthcare, including applications like decision-making support systems.