Large language models (LLMs) have been widely studied for their ability to store and utilize positive knowledge. However, negative knowledge, such as "lions don't live in the ocean", is also ubiquitous in the world but rarely mentioned explicitly in the text. What do LLMs know about negative knowledge? This work examines the ability of LLMs to negative commonsense knowledge. We design a constrained keywords-to-sentence generation task (CG) and a Boolean question-answering task (QA) to probe LLMs. Our experiments reveal that LLMs frequently fail to generate valid sentences grounded in negative commonsense knowledge, yet they can correctly answer polar yes-or-no questions. We term this phenomenon the belief conflict of LLMs. Our further analysis shows that statistical shortcuts and negation reporting bias from language modeling pre-training cause this conflict.
Mohamed bin Zayed University of Artificial Intelligence(莫扎伊德·本·泽德人工智能大学);SDAIA;Al-Balqa Applied University(巴尔加应用大学);Khalifa University(卡利法大学);HUMAIN, Data and AI Models(HUMAIN、数据与人工智能模型)
$R^3$: Reverse, Retrieve, and Rank for Sarcasm Generation with Commonsense Knowledge
Information Sciences Institute, University of Southern California(南加州大学信息科学研究所);Department of Computer Science, Columbia University(哥伦比亚大学计算机科学系);Educational Testing Service(教育考试服务中心);Data Science Institute, Columbia University(哥伦比亚大学数据科学研究所)
How to Motivate Your Dragon: Teaching Goal-Driven Agents to Speak and Act in Fantasy Worlds
Georgia Institute of Technology(佐治亚理工学院);Facebook AI Research(脸书人工智能研究院)
我们寻求创造代理人,既采取行动,并与其他代理人在追求一个目标。为此,我们扩展了LIGHT(Urbanek et al. 2019)-一个大规模的众包幻想文本游戏-带有任务数据集。这些包含自然语言动机,以及游戏中的目标和人类演示;完成任务可能需要对话或行动(或两者兼而有之)。我们引入了一个强化学习系统,该系统(1)结合了基于大规模语言建模和基于常识推理的预训练,以向智能体注入相关的先验知识;(2)利用动作命令和对话的因子化动作空间,在两者之间进行平衡。我们使用人类专家演示进行zero-shot评估,表明我们的代理能够始终如一地采取行动,并自然地谈论他们的动机。
Breaking Language Barriers: Equitable Performance in Multilingual Language Models
University of Wisconsin–Madison(威斯康星大学麦迪逊分校);NSF Center for Computer Assisted Synthesis (CCAS), University of Notre Dame(国家科学基金会计算机辅助合成中心(CCAS)、诺特大学);University of Notre Dame(诺特大学)