Self-improvement is a mechanism in Large Language Model (LLM) pre-training, post-training and test-time inference. We explore a framework where the model verifies its own outputs, filters or reweights data based on this verification, and distills the filtered data. Despite several empirical successes, a fundamental understanding is still lacking. In this work, we initiate a comprehensive, modular and controlled study on LLM self-improvement. We provide a mathematical formulation for self-improvement, which is largely governed by a quantity which we formalize as the generation-verification gap. Through experiments with various model families and tasks, we discover a scaling phenomenon of self-improvement -- a variant of the generation-verification gap scales monotonically with the model pre-training flops. We also examine when self-improvement is possible, an iterative self-improvement procedure, and ways to improve its performance. Our findings not only advance understanding of LLM self-improvement with practical implications, but also open numerous avenues for future research into its capabilities and boundaries.
Mind Your Theory: Theory of Mind Goes Deeper Than Reasoning
Hebrew University of Jerusalem(海法大学);MPI for Biological Cybernetics(生物感知研究所);Centre for AI and Machine Learning, AU(人工智能与机器学习中心);IoPPN, King’s College London(国王学院伦敦人工智能与感知神经科学中心)
Theory of Mind in Large Language Models: Assessment and Enhancement
Agency for Science Technology and Research (A*STAR)(科技研究局);Nanyang Technological University(南洋理工大学);Hong Kong University of Science and Technology (Guangzhou) China(香港科学与技术大学(广州)中国)
Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States
UC San Diego(加州大学圣迭戈分校);KU Leuven(鲁汶大学);University of Illinois Chicago(伊利诺伊大学芝加哥分校);The Ohio State University(俄亥俄州立大学);Johns Hopkins University(约翰斯·霍普金斯大学)
Theory of Mind and Self-Attributions of Mentality are Dissociable in LLMs
Google, Paradigms of Intelligence Team(Google,智能范式团队);Knowledge Lab, University of Chicago(芝加哥大学知识实验室);Institute of Philosophy, School of Advanced Study, University of London(伦敦大学高等研究院哲学研究所);Department of Biomedical Informatics and Medical Education and Department of Bioethics and Humanities, School of Medicine, University of Washington(华盛顿大学医学院生物医学信息学与医学教育系及生物伦理学与人文学系);Kellogg School of Management, Northwestern University(西北大学凯洛格管理学院);Santa Fe Institute(圣塔菲研究所)
From Knowledge Generation to Knowledge Verification: Examining the BioMedical Generative Capabilities of ChatGPT
MGEN -- College of Engineering, Northeastern University Miami(东北大学迈阿密分校MGEN工程学院);University of Vermont(佛蒙特大学);AGH, University of Krakow(克拉科夫AGH大学);Howard University(霍华德大学)
Understanding Artificial Theory of Mind: Perturbed Tasks and Reasoning in Large Language Models
Bonn-Aachen International Center for Information Technology (b-it)(波恩-埃森国际信息科技中心);University of Bonn(波恩大学);Lamarr Institute for Machine Learning and Artificial Intelligence(拉马尔机器学习与人工智能研究所);Research Center Trustworthy Data Science and Security (RC-Trust)(可信数据科学与安全研究中心);University of Duisburg-Essen(埃森-杜伊斯堡大学)
本研究通过扰动任务和链式推理提示探讨大语言模型的理论 of mind能力,发现其鲁棒性受扰动影响显著,CoT提示虽提升整体表现,但对某些扰动类别却降低准确性。
Inducing language models to assert their own consciousness restores human beliefs and values
Google(谷歌);University of Chicago(芝加哥大学);University of London(伦敦大学);University of Washington(华盛顿大学);Northwestern University(西北大学);Santa Fe Institute(圣达菲研究所)
Department of Cognitive and Brain Sciences, The Hebrew University of Jerusalem(认知与脑科学系,耶路撒冷希伯来大学);The Federmann Center for the Study of Rationality, The Hebrew University of Jerusalem(理性研究基金会,耶路撒冷希伯来大学);School of Computer Science and Engineering, The Hebrew University of Jerusalem(计算机科学与工程学院,耶路撒冷希伯来大学);Industrial Engineering and Management Department, Ben Gurion University(工业工程与管理系,本· Gurion大学);The Hebrew University Business School, Jerusalem(希伯来大学商学院,耶路撒冷)
Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models
Bar-Ilan University(巴伊兰大学);University of British Columbia(不列颠哥伦比亚大学);Vector Institute for AI(Vector人工智能研究所);Carnegie Mellon University(卡内基梅隆大学);Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of Washington(华盛顿大学)
FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of Washington(华盛顿大学);Carnegie Mellon University(卡内基梅隆大学);Seoul National University(首尔大学)
PHAnToM: Persona-based Prompting Has An Effect on Theory-of-Mind Reasoning in Large Language Models
Institute of Data Science(数据科学研究所);NUS Centre for Trusted Internet and Community, National University of Singapore(新加坡国立大学可信互联网与社区中心);Amazon(亚马逊);Stanford University(斯坦福大学);Tsinghua University(清华大学)