Tweet by Xiang Yue on LLM deficiencies in reasoning

qL-KQ-Al_normal.jpg spacer.png
Xiang Yue
⁦‪@xiangyue96‬⁩
logo_twitter-1497383721365.png
spacer_464x1-1582829598167.png
Our #EMNLP2023 work also reveals this phenomenon. We find that despite being able to generate correct step-by-step solutions in the beginning, LLMs cannot maintain their belief in truth when challenged by often-time absurdly invalid arguments.

browse.arxiv.org/pdf/2305.13160… pic.twitter.com/3MRykrXmYC

10/8/23, 4:54 PM

Joseph Thornton

Leave a comment