We introduce a method to measure uncertainty in large language models. For tasks like question answering, it is essential to know when we can trust the natural language outputs of foundation models. We show that measuring uncertainty in natural language is challenging because of "semantic equivalence" -- different sentences can mean the same thing. To overcome these challenges we introduce semantic entropy -- an entropy which incorporates linguistic invariances created by shared meanings. Our method is unsupervised, uses only a single model, and requires no modifications to off-the-shelf language models. In comprehensive ablation studies we show that the semantic entropy is more predictive of model accuracy on question answering data sets than comparable baselines.
Anthropomimetic Uncertainty: What Verbalized Uncertainty in Language Models is Missing
University of Amsterdam(阿姆斯特丹大学);University of Groningen(格罗宁根大学);University of Edinburgh(爱丁堡大学);IT University of Copenhagen(哥本哈根信息技术大学);Pioneer Centre for Artificial Intelligence(先锋人工智能中心)
Can Large Language Models Express Uncertainty Like Human?
School of Computer Science, University of Sydney(悉尼大学计算机科学学院);City University of Hong Kong(香港城市大学);Shanghai Jiao Tong University(上海交通大学);Department of Engineering Science, University of Oxford(牛津大学工程科学系)
Rethinking Prospect Theory for LLMs: Revealing the Instability of Decision-Making under Epistemic Uncertainty
Hong Kong University of Science and Technology(香港科技大学);Huazhong University of Science and Technology(华中科技大学);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)
Revisiting Uncertainty Estimation and Calibration of Large Language Models
School of Computer Science University of Sydney(悉尼大学计算机科学学院);City University of Hong Kong(香港城市大学);Shanghai Jiao Tong University(上海交通大学);Department of Engineering Science University of Oxford(牛津大学工程科学系)
Uncertainty in Natural Language Generation: From Theory to Applications
University of Amsterdam(阿姆斯特丹大学);TU Darmstadt(达姆施塔特工业大学);IT University of Copenhagen(哥本哈根信息技术大学);Pioneer Centre for Artificial Intelligence(先锋人工智能中心);Instituto de Telecomunicações(电信研究所);Instituto Superior Técnico(高等技术学院);LUMLIS (Lisbon ELLIS Unit)(里斯本ELLIS单元LUMLIS);University of Zurich(苏黎世大学);University of Edinburgh(爱丁堡大学);LMU Munich(慕尼黑大学);Munich Center for Machine Learning(慕尼黑机器学习中心)
The challenge of uncertainty quantification of large language models in medicine
Ontario Tech University(安大略理工大学);Icahn School of Medicine at Mount Sinai(西奈山伊坎医学院);Iran University of Medical Sciences(伊朗医科大学);Stanford University(斯坦福大学);The Charles Bronfman Institute of Personalized Medicine, Icahn School of Medicine at Mount Sinai(西奈山伊坎医学院查尔斯·布朗夫曼个性化医学研究所);Henry D. Janowitz Division of Gastroenterology, Icahn School of Medicine at Mount Sinai(西奈山伊坎医学院亨利·D·亚诺维茨胃肠病学部门)
The grip of grammar on meaning uncertainty: cross-linguistic evidence, neural correlates, and clinical relevance
Grammar and Cognition Lab, Department of Translation & Language Sciences, Universitat Pompeu Fabra(语法与认知实验室,翻译与语言科学系,庞培法布拉大学);School of Foreign Studies, Guangzhou University(外语学院,广州大学);Department of Neurology Sant Pau Memory Unit, Hospital de la Santa Creu i Sant Pau Biomedical Research Institute Sant Pau(圣帕乌记忆单位神经科,圣帕乌生物医学研究机构);Frontotemporal Disorders Unit, Department of Neurology, Massachusetts General Hospital, Harvard Medical School(前额叶疾病单位,神经科,马萨诸塞总医院,哈佛医学院);Ace Alzheimer Center Barcelona, Universitat Internacional de Catalunya(巴塞罗那阿尔茨海默病中心,国际加泰罗尼亚大学);Networking Research Center on Neurodegenerative Diseases (CIBERNED), Instituto de Salud Carlos III(神经退行性疾病网络研究机构(CIBERNED),卡洛斯三世健康研究所);Department of Language Science and Technology, Saarland University(语言科学与技术系,萨尔兰大学);Department of Operation and Management, Zhuhai People's Hospital (The Affiliated Hospital of Beijing Institute of Technology, Zhuhai Clinical Medical College of Jinan University)(珠海人民医院(北京理工大学附属珠海临床医学院));Department of Statistics, Zhuhai People's Hospital (The Affiliated Hospital of Beijing Institute of Technology, Zhuhai Clinical Medical College of Jinan University)(珠海人民医院(北京理工大学附属珠海临床医学院));Faculty of Medicine, Macau University of Science and Technology(医学学院,澳门科技大学);CIDCL, Escuela de Fonoaudiología, Universidad de Valparaíso(CIDCL,语言治疗学系,瓦尔帕莱索大学);Department of Psychiatry and Psychotherapy, University of Marburg(精神病学与心理学系,马尔堡大学);Marburg University, School of Medicine, Department of Psychiatry and Psychotherapy(马尔堡大学,医学院,精神病学与心理学系);Department of Adult Psychiatry and Psychotherapy, University of Zurich(成人精神病学与心理学系,苏黎世大学);Neuroscience Center Zurich, University of Zurich and ETH Zurich(苏黎世神经科学中心,苏黎世大学和ETH苏黎世联邦理工学院);Center for Clinical Neuroscience and Cognition and Department of Psychiatry, University of Groningen, University Medical Center Groningen(临床神经科学与认知中心及精神病学系,格罗宁根大学,格罗宁根大学医学中心);Douglas Mental Health University Institute, Department of Psychiatry, McGill University(道格拉斯大学心理健康研究所,精神病学系,麦吉尔大学);Department of Medical Biophysics, Schulich School of Medicine and Dentistry, Western University(医学生物物理系,施吕赫尔医学院及牙科学院,西部大学);Robarts Research Institute, Schulich School of Medicine and Dentistry, Western University(罗巴茨研究所,施吕赫尔医学院及牙科学院,西部大学);Institut Català de Recerca i Estudis Avançats (ICREA), Barcelona, Spain(加泰罗尼亚高级研究与研究机构(ICREA),巴塞罗那,西班牙)
Curiosity as Linguistic Intervention: Using LLM Tutoring Dialogues to Influence Exploratory Learning Behavior
School of Computing and Information Systems, Singapore Management University(新加坡管理大学计算与信息系统学院);Human-Computer Interaction Institute, Carnegie Mellon University(卡内基梅隆大学人机交互研究所)
Provably Secure Disambiguating Neural Linguistic Steganography
University of Science and Technology of China(中国科学技术大学);School of Cyber Science and Technology(网络空间安全学院);Anhui Province Key Laboratory of Digital Security(安徽省数字安全重点实验室)
ALIEN: Aligned Entropy Head for Improving Uncertainty Estimation of LLMs
Applied AI Institute, Moscow, Russia(应用人工智能研究所,莫斯科,俄罗斯);SB AI Lab, Moscow, Russia(SB AI实验室,莫斯科,俄罗斯);Intellectual data analysis and predictive modeling Institute, Moscow, Russia(智能数据分析与预测建模研究所,莫斯科,俄罗斯)