Vol. 2 No. 6 (2025): LLM III - Limitations - Advances

					View Vol. 2 No. 6 (2025): LLM III - Limitations - Advances

We continue to monitor progress in large language models emphasizing safety & value alignment. 

Published: 2025-09-23

Articles

  • Highlights of the Issue - Large Language Models III Kris Carlson, Publisher

    Kris Carlson
    DOI: https://doi.org/10.70777/si.v2i6.15969
  • International Al Safety Report: First Key Update Capabilities and Risk Implications

    Yoshua Bengio, Benjamin Bucknall, Stephen Clare, Carina Prunkl, Maksym Andriushchenko, Philip Fox, Tiancheng Hu, Cameron Jones, Sam Manning, Nestor Maslej, Vasilios Mavroudis, Conor McGlynn, Malcolm Murray, Shalaleh Rismani, Charlotte Stix, Lucia Velasco, Nicole Wheeler, Daniel Privitera, Sören Mindermann, Daron Acemoglu, Thomas G. Dietterich, Fredrik Heintz, Geoffrey Hinton, Nick Jennings, Susan Leavy, Teresa Ludermir, Vidushi Marda, Helen Margetts, John McDermid, Jane Munga, Arvind Narayanan, Alondra Nelson, Clara Neppel, Sarvapali D. (Gopal) Ramchurn, Stuart Russell, Marietje Schaake, Bernhard Schölkopf, Alvaro Soto, Lee Tiedrich, Gaël Varoquaux, Andrew Yao, Ya-Qin Zhan
    DOI: https://doi.org/10.70777/si.v2i6.16253
  • Responsible Agentic Reasoning and AI Agents: A Critical Survey Proposal for Safe Agentic AI via Responsible Reasoning AI Agents (R2A2)

    Shaina Raza, Ranjan Sapkota, Manoj Karkee, Christos Emmanouilidis
    DOI: https://doi.org/10.70777/si.v2i6.16169
  • The Illusion of Thinking Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity

    Parshin Shojaee , Iman Mirzadeh, Keivan Alizadeh, Maxwell Horton, Samy Bengio, Mehrdad Farajtabar
    DOI: https://doi.org/10.70777/si.v2i6.15919
  • Thinking Isn’t an Illusion Overcoming the Limitations of Reasoning Models via Tool Augmentations

    Zhao Song, Song Yue, Jiahao Zhang
    DOI: https://doi.org/10.70777/si.v2i6.15961
  • The Asymptotic Intelligence Thesis: Rethinking the Ceiling of AGI Cognition

    Jeffrey E. Arle, MD, PhD, FAANS, FCNS
    DOI: https://doi.org/10.70777/si.v2i6.16255
  • From Hard Refusals to Safe-Completions: Toward Output-Centric Safety Training

    Yuan Yuan, Tina Sriskandarajah, Anna-Luisa Brakman, Alec Helyar, Alex Beutel, Andrea Vallone, Saachi Jain
    DOI: https://doi.org/10.70777/si.v2i6.15625
  • Precedents for the Unprecedented: Historical Analogies for Thirteen Artificial Superintelligence Risks

    James D. Miller
    DOI: https://doi.org/10.70777/si.v2i6.16999
  • Enabling Frontier Lab Collaboration to Mitigate AI Safety Risks

    Nicholas Felstead
    DOI: https://doi.org/10.70777/si.v2i6.16439

Reports

  • Understanding Limitations of Large Language Models from First Principles Computational Complexity Circuit Class TCk

    Kris Carlson
    DOI: https://doi.org/10.70777/si.v2i6.16549

Reviews

Commentary

  • Evidence Integrity Before Capability: A Prerequisite for Safe Artificial Intelligence

    Jennifer Flygare Kinne
    DOI: https://doi.org/10.70777/si.v2i6.16393
  • Timeline to Artificial General Intelligence 2025 – 2030+ A prediction of how AI will progress, year by year. Updated Oct 30, 2025.

    Gil Syswerda
    DOI: https://doi.org/10.70777/si.v2i6.16375