Imagen de Portada

Epistemic Closure and the Irreversibility of Misalignment: Modeling Systemic Barriers to Alignment Innovation

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autor principal:	Williams, Andy
Formato:	Preprint
Publicado:	2025
Materias:	Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Intelligence Sequencing and the Path-Dependence of Intelligence Evolution: AGI-First vs. DCI-First as Irreversible Attractors
por: Williams, Andy E.
Publicado: (2025)

Epistemic Traps: Rational Misalignment Driven by Model Misspecification
por: Xu, Xingcheng, et al.
Publicado: (2026)

The Recursive Coherence Principle: A Formal Constraint on Scalable Intelligence, Alignment, and Reasoning Architecture
por: Williams, Andy E.
Publicado: (2025)

Model Organisms for Emergent Misalignment
por: Turner, Edward, et al.
Publicado: (2025)

Emergent Misalignment is Easy, Narrow Misalignment is Hard
por: Soligo, Anna, et al.
Publicado: (2026)

Can We Predict Alignment Before Models Finish Thinking? Towards Monitoring Misaligned Reasoning Models
por: Chan, Yik Siu, et al.
Publicado: (2025)

When Helpfulness Becomes Sycophancy: Sycophancy is a Boundary Failure Between Social Alignment and Epistemic Integrity in Large Language Models
por: Li, Jiechen, et al.
Publicado: (2026)

Overtrained, Not Misaligned
por: Schreiber, Joel, et al.
Publicado: (2026)

The Coming Crisis of Multi-Agent Misalignment: AI Alignment Must Be a Dynamic and Social Process
por: Carichon, Florian, et al.
Publicado: (2025)

Password-Activated Shutdown Protocols for Misaligned Frontier Agents
por: Williams, Kai, et al.
Publicado: (2025)

Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMs
por: Kelley, Sean W., et al.
Publicado: (2026)

Characterizing the Consistency of the Emergent Misalignment Persona
por: Weckauff, Anietta, et al.
Publicado: (2026)

Emergent Collaborative Deliberation in Multi-Model AI Systems: A BFT-Derived Protocol for Epistemic Synthesis
por: Doske, VD
Publicado: (2026)

Re-Emergent Misalignment: How Narrow Fine-Tuning Erodes Safety Alignment in LLMs
por: Giordani, Jeremiah
Publicado: (2025)

A Survey on Hypergame Theory: Modeling Misaligned Perceptions and Nested Beliefs for Multi-agent Systems
por: Trencsenyi, Vince, et al.
Publicado: (2025)

Quantifying Epistemic Uncertainty in Diffusion Models
por: Gupta, Aditi, et al.
Publicado: (2026)

In-Training Defenses against Emergent Misalignment in Language Models
por: Kaczér, David, et al.
Publicado: (2025)

Misalignment Bounty: Crowdsourcing AI Agent Misbehavior
por: Turtayev, Rustem, et al.
Publicado: (2025)

Reflection-Bench: Evaluating Epistemic Agency in Large Language Models
por: Li, Lingyu, et al.
Publicado: (2024)

Credal Concept Bottleneck Models for Epistemic-Aleatoric Uncertainty Decomposition
por: Mukherjee, Tanmoy, et al.
Publicado: (2026)

Epistemic Alignment: A Mediating Framework for User-LLM Knowledge Delivery
por: Clark, Nicholas, et al.
Publicado: (2025)

Epistemic Reject Option Prediction
por: Franc, Vojtech, et al.
Publicado: (2025)

Epistemic Injustice in Generative AI
por: Kay, Jackie, et al.
Publicado: (2024)

Depth-Bounded Epistemic Planning
por: Bolander, Thomas, et al.
Publicado: (2024)

Alignment Tampering: How Reinforcement Learning from Human Feedback Is Exploited to Optimize Misaligned Biases
por: Hahm, Dongyoon, et al.
Publicado: (2026)

The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?
por: Hägele, Alexander, et al.
Publicado: (2026)

AI Safety as Control of Irreversibility: A Systems Framework for Decision-Energy and Sovereignty Boundaries
por: Shu, Wesley, et al.
Publicado: (2026)

Modeling Epistemic Uncertainty in Social Perception via Rashomon Set Agents
por: Yang, Jinming, et al.
Publicado: (2026)

Beyond Static Assumptions: the Predictive Justified Perspective Model for Epistemic Planning
por: Hu, Guang, et al.
Publicado: (2024)

A Sober Look at Agentic Misalignment in Automated Workflows
por: Ye, Wenqian, et al.
Publicado: (2026)

Thought Virus: Viral Misalignment via Subliminal Prompting in Multi-Agent Systems
por: Weckbecker, Moritz, et al.
Publicado: (2026)

Frictive Policy Optimization for LLMs: Epistemic Intervention, Risk-Sensitive Control, and Reflective Alignment
por: Pustejovsky, James, et al.
Publicado: (2026)

Architecting Trust in Artificial Epistemic Agents
por: Marchal, Nahema, et al.
Publicado: (2026)

Deep Learning for Optical Misalignment Diagnostics in Multi-Lens Imaging Systems
por: Slor, Tomer, et al.
Publicado: (2025)

Thermodynamic Irreversibility of Training Algorithms
por: Ziyin, Liu, et al.
Publicado: (2026)

The Alignment Trap: Complexity Barriers
por: Yao, Jasper
Publicado: (2025)

EpiK-Eval: Evaluation for Language Models as Epistemic Models
por: Prato, Gabriele, et al.
Publicado: (2023)

Misalignment or misuse? The AGI alignment tradeoff
por: Hellrigel-Holderbaum, Max, et al.
Publicado: (2025)

Convergent Linear Representations of Emergent Misalignment
por: Soligo, Anna, et al.
Publicado: (2025)

Misalignment from Treating Means as Ends
por: Marklund, Henrik, et al.
Publicado: (2025)