提出自动化元提示与 LLMaaJ 强化学习框架，通过建模人类编辑和 Hilbert 空间内容特征实现 ToM 对齐，并在美网等现场内容生产中验证效果。
摘要：
我们提出一种元提示方法，可在为复杂任务生成流畅文本的同时，优化人类心理预期与大语言模型（LLM）神经处理之间神经状态的相似性。该方法采用智能体强化学习技术：由作为裁判的 LLM（LLM as a Judge，LLMaaJ）通过上下文学习教导另一个 LLM，使其通过解释生成文本中有意与无意的特征来产出内容。为衡量人类在内容生产方面的心理信念，用户在 2024 年美国网球公开赛大满贯赛事发布前修改 AI 生成的长篇文章。现在，LLMaaJ 能够通过在 LLM 生成文本的过程中预测并纳入人类编辑，解决心智理论（Theory of Mind，ToM）对齐问题。在整个实验中，并通过解释一个实时生产系统的结果，人类内容审核员的预期与 AI 在 53.8% 的情况下达到 100% 对齐，平均迭代次数为 4.38。在 Hilbert 向量空间上对事实性、新颖性、重复性和相关性等内容特征进行几何解释，将空间体积（所有特征的重要性）与顶点对齐（单个特征的相关性）相结合，使 LLMaaJ 能够针对人类 ToM 进行优化。这通过扩大网球动作的覆盖范围提升了内容质量。我们的工作已部署于 2024 年美国网球公开赛，并被用于体育和娱乐领域的其他现场活动。
英文摘要：
We introduce a method of meta-prompting that jointly produces fluent text for complex tasks while optimizing the similarity of neural states between a human's mental expectation and a Large Language Model's (LLM) neural processing. A technique of agentic reinforcement learning is applied, in which an LLM as a Judge (LLMaaJ) teaches another LLM, through in-context learning, how to produce content by interpreting the intended and unintended generated text traits. To measure human mental beliefs around content production, users modify long form AI-generated text articles before publication at the US Open 2024 tennis Grand Slam. Now, an LLMaaJ can solve the Theory of Mind (ToM) alignment problem by anticipating and including human edits within the creation of text from an LLM. Throughout experimentation and by interpreting the results of a live production system, the expectations of human content reviewers had 100% of alignment with AI 53.8% of the time with an average iteration count of 4.38. The geometric interpretation of content traits such as factualness, novelty, repetitiveness, and relevancy over a Hilbert vector space combines spatial volume (all trait importance) with vertices alignment (individual trait relevance) enabled the LLMaaJ to optimize on Human ToM. This resulted in an increase in content quality by extending the coverage of tennis action. Our work that was deployed at the US Open 2024 has been used across other live events within sports and entertainment.
ToMBench: Benchmarking Theory of Mind in Large Language Models
Tsinghua University（清华大学）；The Chinese University of Hong Kong（香港中文大学）；IIE, CAS（中国科学院自动化研究所）；Tianjin University（天津大学）；Northwest Minzu University（西北民族大学）；Nankai University（南开大学）；Beijing Institute of Technology（北京理工大学）
Mind the Style Gap: Meta-Evaluation of Style and Attribute Transfer Metrics
Department of Computer Science, Aarhus University, Denmark（计算机科学系，奥胡斯大学）；Department of Computer Science, University of Copenhagen, Denmark（计算机科学系，哥本哈根大学）
I Cast Detect Thoughts: Learning to Converse and Guide with Intents and Theory-of-Mind in Dungeons and Dragons
Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；University of Southern California（南加州大学）；University of Washington（华盛顿大学）；University of Pennsylvania（宾夕法尼亚大学）；MIT（麻省理工学院）
Mind the Gap: Towards Generalizable Autonomous Penetration Testing via Domain Randomization and Meta-Reinforcement Learning
College of Electronic Engineering, National University of Defense Technology（国防科技大学电子工程学院）；Beijing National Research Center for Information Science and Technology, Tsinghua University（清华大学北京信息科学与技术国家研究中心）；College of Computer Science and Technology, National University of Defense Technology（国防科技大学计算机学院）
Princeton AI Lab（普林斯顿人工智能实验室）；Princeton University（普林斯顿大学）；Institute for Human-Centered AI（以人为中心的人工智能研究所）；Helmholtz Munich（海德堡-慕尼黑研究所）；Ludwig-Maximilians-University（路德维希-马克西米利安大学）
Toward Automated Algorithm Design: A Survey and Practical Guide to Meta-Black-Box-Optimization
School of Computer Science and Engineering, South China University of Technology（计算机科学与工程学院，华南理工大学）；Nankai University（南开大学）；Hanyang University（翰阳大学）；Hong Kong Polytechnic University（香港理工大学）
Manalyzer: End-to-end Automated Meta-analysis with Multi-agent System
Shanghai Jiao Tong University（上海交通大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；The Chinese University of Hong Kong（香港中文大学）；Nankai University（南开大学）；Peking University（北京大学）
MetaScientist: A Human-AI Synergistic Framework for Automated Mechanical Metamaterial Design
University of Pennsylvania（宾夕法尼亚大学）；Virginia Tech（弗吉尼亚理工大学）；University of California, Los Angeles（加利福尼亚大学洛杉矶分校）；University of California, Davis（加利福尼亚大学戴维斯分校）；Princeton University（普林斯顿大学）
