Imagen de Portada

AI Behind Closed Doors: a Primer on The Governance of Internal Deployment

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Stix, Charlotte, Pistillo, Matteo, Sastry, Girish, Hobbhahn, Marius, Ortega, Alejandro, Balesni, Mikita, Hallensleben, Annika, Goldowsky-Dill, Nix, Sharkey, Lee
Formato:	Preprint
Publicado:	2025
Materias:	Computers and Society
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

The Loss of Control Playbook: Degrees, Dynamics, and Preparedness
por: Stix, Charlotte, et al.
Publicado: (2025)

Pre-Deployment Information Sharing: A Zoning Taxonomy for Precursory Capabilities
por: Pistillo, Matteo, et al.
Publicado: (2024)

Large Language Models can Strategically Deceive their Users when Put Under Pressure
por: Scheurer, Jérémy, et al.
Publicado: (2023)

Assurance of Frontier AI Built for National Security
por: Pistillo, Matteo, et al.
Publicado: (2025)

Detecting Strategic Deception Using Linear Probes
por: Goldowsky-Dill, Nicholas, et al.
Publicado: (2025)

Internal Deployment in the AI Act
por: Pistillo, Matteo
Publicado: (2025)

Towards evaluations-based safety cases for AI scheming
por: Balesni, Mikita, et al.
Publicado: (2024)

Frontier Models are Capable of In-context Scheming
por: Meinke, Alexander, et al.
Publicado: (2024)

Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
por: Braun, Dan, et al.
Publicado: (2024)

Lessons from Studying Two-Hop Latent Reasoning
por: Balesni, Mikita, et al.
Publicado: (2024)

Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs
por: Laine, Rudolf, et al.
Publicado: (2024)

Using Degeneracy in the Loss Landscape for Mechanistic Interpretability
por: Bushnaq, Lucius, et al.
Publicado: (2024)

How to evaluate control measures for LLM agents? A trajectory from today to superintelligence
por: Korbak, Tomek, et al.
Publicado: (2025)

Stress Testing Deliberative Alignment for Anti-Scheming Training
por: Schoen, Bronson, et al.
Publicado: (2025)

The Local Interaction Basis: Identifying Computationally-Relevant and Sparsely Interacting Features in Neural Networks
por: Bushnaq, Lucius, et al.
Publicado: (2024)

Children in Police Custody: Adversity and Adversariality Behind Closed Doors
por: Frances Sheahan
Publicado: (2025)

Towards Frontier Safety Policies Plus
por: Pistillo, Matteo
Publicado: (2025)

Behind Office Doors
Publicado: (2026)

Behind Office Doors
Publicado: (2026)

Peeking Behind Closed Doors: Risks of LLM Evaluation by Private Data Curators
por: Bansal, Hritik, et al.
Publicado: (2025)

Defending Compute Thresholds Against Legal Loopholes
por: Pistillo, Matteo, et al.
Publicado: (2025)

Honesty to Subterfuge: In-Context Reinforcement Learning Can Make Honest Models Reward Hack
por: McKee-Reid, Leo, et al.
Publicado: (2024)

Chapter 6 The Role of Corporate Governance in Macro-Prudential Regulation of Systemic Risk
por: Dill, Alexander
Publicado: (2020)

Owning the Stuff of Life
por: Stix, Gary

The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
por: Berglund, Lukas, et al.
Publicado: (2023)

Behind Closed Doors: An Exploratory Study of the Perceptions of Librarians and the Hidden Intellectual Work of Collection Development in Canadian Public Libraries.
por: Nilsen, Kirsti, et al.
Publicado: (2002)

Ground states for the Hartree energy functional in the critical case
por: Pistillo, Tommaso
Publicado: (2025)

Backchaining Loss of Control Mitigations from Mission-Specific Benchmarks in National Security
por: Pistillo, Matteo, et al.
Publicado: (2026)

Analogical Reasoning Within a Conceptual Hyperspace
por: Goldowsky, Howard, et al.
Publicado: (2024)

Technical Report: Evaluating Goal Drift in Language Model Agents
por: Arike, Rauno, et al.
Publicado: (2025)

Forecasting Frontier Language Model Agent Capabilities
por: Pimpale, Govind, et al.
Publicado: (2025)

The Day the Library Closed Its Doors
por: Yates, Elizabeth
Publicado: (1970)

Hunter Midtown Library: The Closing of an Open Door
por: Foster, Barbara
Publicado: (1976)

DoorBot: Closed-Loop Task Planning and Manipulation for Door Opening in the Wild with Haptic Feedback
por: Wang, Zhi, et al.
Publicado: (2025)

Bibliophilately Revisited.
por: Nix, Larry T.
Publicado: (2000)

A Study of the Bookmobile Service of the Madison Public Library.
por: Nix, Larry T.
Publicado: (1981)

The étale topos reconstructs varieties over sub-p-adic fields
por: Carlson, Magnus, et al.
Publicado: (2024)

Monographs in Microform: Issues in Cataloging and Bibliographic Control.
por: Mikita, Elizabeth G.
Publicado: (1981)

Large Language Models Often Know When They Are Being Evaluated
por: Needham, Joe, et al.
Publicado: (2025)

Analyzing Probabilistic Methods for Evaluating Agent Capabilities
por: Højmark, Axel, et al.
Publicado: (2024)