本頁只刊出中文翻譯與中文說明；英文原文請見下方原文連結。
原文連結
論文資訊
- 類型：已發表論文
- 日期：2023-03-21
摘要
We survey a current, heated debate in the 人工智慧 (AI) research community on whether large pretrained 語言 models can be said to understand 語言-and the physical and 社會 situations 語言 encodes-in any humanlike sense. We describe argu-ments that have been made for and against such under-standing and key questions for the broader sciences of intelligence that have arisen in light of these arguments. We contend that an extended science of intelligence can be developed that will provide insight into distinct modes of understanding, their strengths and limitations, and the challenge of integrating diverse forms of cognition.
※ 此為已發表論文，全文需透過期刊付費取得