相关论文: How Well Do Models Follow Their Constitutions?

Large language models (LLMs) learn statistical associations from massive training corpora and user interactions, and deployed systems can surface or infer information about individuals. Yet people lack practical ways to inspect what a model…

The field of AI alignment is increasingly concerned with the questions of how values are integrated into the design of generative AI systems and how their integration shapes the social consequences of AI. However, existing transparency…

Deep neural networks demonstrate impressive performance in visual recognition, but they remain vulnerable to adversarial attacks that is imperceptible to the human. Although existing defense strategies such as adversarial training and…

Objective: This study develops a systematic benchmarking framework for testing whether language models can accurately identify constructs of interest in child welfare records. The objective is to assess how different model sizes and…

The research creates a professional certification survey to test large language models and evaluate their employable skills. It compares the performance of two AI models, GPT-3 and Turbo-GPT3.5, on a benchmark dataset of 1149 professional…

The deployment of AI systems faces three critical governance challenges that current frameworks fail to adequately address. First, organizations struggle with inadequate risk assessment at the use case level, exemplified by the Humana class…

We conducted controlled experimental bias audits for four versions of ChatGPT, which we asked to recommend an opening offer in salary negotiations for a new hire. We submitted 98,800 prompts to each version, systematically varying the…

Multi-agent LLM ensembles can converge on coordinated, socially harmful equilibria. This paper advances an experimental framework for evaluating Institutional AI, our system-level approach to AI alignment that reframes alignment from…

The 2025 BEHAVIOR Challenge is designed to rigorously track progress toward solving long-horizon tasks by physical agents in simulated environments. BEHAVIOR-1K focuses on everyday household tasks that people most want robots to assist with…

AI models are already deployed in societies affected by armed conflict, and journalists, humanitarian workers, governments and ordinary citizens rely on them for information or for their work processes. No established practice exists for…

Reinforcement-learned reasoning has powered recent AI leaps on verifiable tasks, including mathematics, code, and structure prediction. The harder bottleneck is evaluative judgment in low-verifiability domains, where no oracle anchors…

Large language models (LLMs), initially developed for generative AI, are now evolving into agentic AI systems, which make decisions in complex, real-world contexts. Unfortunately, while their generative capabilities are well-documented,…

Many users interact with AI tools like ChatGPT using a mental model that treats the system as human-like, which we call Model H. According to goal-setting theory, increased specificity in goals should reduce performance variance. If Model H…

Large language models (LLMs) are increasingly trained from AI constitutions and model specifications that establish behavioral guidelines and ethical principles. However, these specifications face critical challenges, including internal…

Introduction: As system dynamics (SD) embraces automation, AI offers efficiency but risks bias from missing data and flawed models. Models that omit multiple perspectives and data threaten model quality, whether created by humans or with…

As artificial intelligence (AI) systems increasingly impact society, the EU Artificial Intelligence Act (AIA) is the first serious legislative attempt to contain the harmful effects of AI systems. This paper proposes a governance framework…

Foundation models - models trained on broad data that can be adapted to a wide range of downstream tasks - can pose significant risks, ranging from intimate image abuse, cyberattacks, to bioterrorism. To reduce these risks, policymakers are…

Adversarial training (AT) and its variants have spearheaded progress in improving neural network robustness to adversarial perturbations and common corruptions in the last few years. Algorithm design of AT and its variants are focused on…

Agentic AI systems increasingly act through tools, sub-agents, and external services, but governance controls are still commonly attached to prompts, dashboards, or post-hoc documentation. This creates a structural mismatch in regulated…

This study systematically evaluates 27 frontier Large Language Models on eight biology benchmarks spanning molecular biology, genetics, cloning, virology, and biosecurity. Models from major AI developers released between November 2022 and…