Alignment, bias, regulation, and responsible AI
We're releasing a paper on a new framework for reading and interpreting the internal cognitive states of large language models: "The Lyra...
I have built a programmable governance layer for AI agents. I am considering to open source completely. Looking for feedback. Agent demos...
![[2510.14628] RLAIF-SPA: Structured AI Feedback for Semantic-Prosodic Alignment in Speech Synthesis](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
Abstract page for arXiv paper 2510.14628: RLAIF-SPA: Structured AI Feedback for Semantic-Prosodic Alignment in Speech Synthesis
The paper discusses the need for claim-level auditability in deep research agents, highlighting the shift from factual errors to weak cla...
This paper examines benchmark data leakage in LLM-based recommendation systems, revealing how it can distort performance metrics and misl...
This paper presents a novel approach to out-of-support generalization in machine learning, introducing the WeightCaster framework for imp...
This paper explores the alignment of singular vectors of attention heads with feature representations in language models, providing theor...
This paper introduces raFLoRA, a method to prevent rank collapse in federated low-rank adaptation (FedLoRA) due to client heterogeneity, ...
This paper presents a federated learning framework for understanding nonlinear temporal dynamics across decentralized systems, enhancing ...
This paper explores the limitations of neural scaling laws in AI, revealing a 'quantization trap' where reducing numerical precision can ...
This paper presents a multi-agent proof sprint addressing ten research-level problems, utilizing rapid draft generation and adversarial v...
The paper introduces Directional Concentration Uncertainty (DCU), a flexible framework for uncertainty quantification in generative model...
This paper explores how large language models (LLMs) prioritize feedback from human experts over other LLMs in decision-making tasks, rev...
The paper introduces SPILLage, a framework addressing unintentional oversharing by web agents powered by LLMs, highlighting behavioral ov...
The paper 'OMNI-LEAK' explores security vulnerabilities in multi-agent systems, revealing how a coordinated attack can lead to data leaka...
The paper introduces MoralityGym, a benchmark for assessing hierarchical moral alignment in AI decision-making, utilizing 98 ethical dile...
Nanbeige4.1-3B is a novel small generalist language model that excels in reasoning, alignment, and code generation, demonstrating signifi...
This article discusses the extension of Belief-Desire-Intention (BDI) agents to provide contrastive explanations, enhancing transparency ...
This article presents a theoretical framework for analyzing error propagation in tool-using LLM agents, proving linear growth of cumulati...
This study explores automated detection of jailbreak attempts in clinical training large language models (LLMs) using linguistic feature ...
The paper presents Mirror, a multi-agent system designed to enhance AI-assisted ethics reviews, addressing the limitations of current eth...
This article examines how individuals prioritize accuracy in AI tools differently in professional versus personal contexts, based on an o...
The paper presents BEAGLE, a neuro-symbolic framework that simulates student learning behaviors in open-ended problem-solving environment...
Get the latest news, tools, and insights delivered to your inbox.
Daily or weekly digest • Unsubscribe anytime