Skip to content

Latest commit

 

History

History
37 lines (26 loc) · 1004 Bytes

2023_CHIP2023_YIER医疗大模型.md

File metadata and controls

37 lines (26 loc) · 1004 Bytes

比赛内容

中文医疗大模型的评估:

  • 模型需要精准理解和合理推演医疗领域的术语和知识

  • 模型须能基于医学公式进行计算和推导,准确把握患者症状,并提供恰如其分的专业临床诊疗知识

比赛地址

https://tianchi.aliyun.com/competition/entrance/532156/information

比赛思路

正常大模型评估的思路

数据集

{
   "context": str,
   "question": str,
   "selection": [str1, str2, str3, str4],
   "answer_choices": [str2, str4],
   "sample_id": str,
   "source": str
}
  • context: 是指医学文本,部分题型无医学文本时请忽略。
  • question: 是指题干问题。
  • selection: 是指题干对应的候选文本项。
  • answer_choices: 答案内容,来自于 selection 的一个或者多个。
  • sample_id: 是指评测团队拟定的题号。
  • source: 是指该题的来源途径,如执业医师真题、临床考验真题、医学专家自拟题。

可联系repo维护者