-
Claude 3.5 Sonnet(2024-10-22),~175B
-
ChatGPT,~175B
-
GPT-4,约1.76T
-
GPT-4o,~200B
-
GPT-4o-mini(gpt-4o-2024-05-13)只有8B
-
最新的o1-mini(o1-mini-2024-09-12)仅100B
-
o1-preview(o1-preview-2024-09-12)~300B
实验部分也是将大模型参数规模分为3挡:7-8B,~100-300B,~1.7T,而GPT-4o-mini被分在第一档,只有8B着实让人有点不可思议~
PromptWizard 概述
https://arxiv.org/pdf/2412.19260v1
MEDEC: A BENCHMARK FOR MEDICAL ERROR DETECTION AND CORRECTION IN CLINICAL NOTES
(文:PaperAgent)