Local Geometry & Policy Properties – Dalcy Ku – PIBBSS Symposium '25

A computational mechanics lens on a transformer’s training – Jasmina Urdshals – PIBBSS Symposium '25

Evolutionary Game Theory and Robustness in RL – Callum Lawson – PIBBSS Symposium '25

The Mechanics of Counting in Transformer LMs – Franz Nowak – PIBBSS Symposium '25

Civilizational Superintelligence Alignment – Max Ramsahoye – PIBBSS Symposium '25

Humans will keep doing hard things (even after AGI) – Alexandre Variengien – PIBBSS Symposium '25

Chain-of-Thought via Reinforcement Learning – Tomasz Steifer – PIBBSS Symposium '25

Automated Safety Research – Jacques Thibodeau – PIBBSS Symposium '25

Security Testing for Multi-Agent Systems – Abhinav Singh – PIBBSS Symposium '25

Simulation Games for AI Agent Interactions – Shray Bansal – PIBBSS Symposium '25

Comp-mech for Logits: Modular Addition Case Study – Xavier Poncini – PIBBSS Symposium '25

Discovery and refinement of mechanisms in parameter space – Antoine Vigouroux – PIBBSS Symposium '25

Plateaus and Valleys in LLM Representations – Matthew Shinkle – PIBBSS Symposium '25

Assistance Games for Unclear Reward Functions – Paul “Lorxus” Rapoport – PIBBSS Symposium '25

Information Markets for Forecasting – Abhimanyu Pallavi Sudhir – PIBBSS Symposium '25

Comparing Human and LLM Social Behavior with D&D – Annie Stephenson – PIBBSS Symposium '25

The geometry of in-context learning - Jan Bauer - PIBBSS Symposium

Cultural Evolution of Cooperation in LLMs - Aron Vallinder - PIBBSS Symposium

Dynamics of LLM beliefs during chain-of-thought reasoning - Baram Sosis - PIBBSS Symposium

Indicators of phenomenal consciousness in LLMs: Metacognition & higher-order theory - Euan McLean

The potential of formal approaches to emergence for AI safety - Nadine Spychala - PIBBSS Symposium

Heavy-tailed Noise & Stochastic Gradient Descent - Wesley Erickson - PIBBSS Symposium

Neuro-Symbolic Approaches for Safe LLM-Based Agents - Agustín Martinez-Suñé - PIBBSS Symposium

Minimum Description Length for singular models - Yevgeny Liokumovich - PIBBSS Symposium

Examining Co-occurence of SAE Features - Matthew A. Clarke - PIBBSS Symposium

Fixing our concepts to understand minds and agency - Mateusz Bagiński - PIBBSS Symposium

Solvable models of in-context learning - Nischal Mainali - PIBBSS Symposium

Factored Space Models: Causality Between Levels of Abstraction - PIBBSS Symposium - Magdalena Wache

Popper on Popper - Rio Popper (PIBBSS Speaker Series)

Explaining emergence in NN with model systems analysis - Ekdeep Singh Lubana (PIBBSS Speaker Series)

AI Alignment w/ Changing and Influenceable Reward Functions - Micah Carroll (PIBBSS Speaker Series)

Autopoiesis and Enaction in the Game of Life - Randall D. Beer (PIBBSS Speaker Series)

Computational Mechanics Hackathon Final Presentations

Computational Mechanics Hackathon '24 Keynote - Paul Riechers and Adam Shai - AI Safety

Information and its Flow - Daniel Polany (PIBBSS Speaker Series) 2024

Causality in Brains - Konrad Körding (PIBBSS Speaker Series) 2024

Creating the contexts for concepts to understand minds (Part 2, Q&A) - Tsvi Benson-Tilsen (PIBBSS)

Creating the contexts for concepts to understand minds - Tsvi Benson-Tilsen (PIBBSS Speakers) 2024

Meaning and Agency - Abram Demski (PIBBSS Speaker Series) 2024

Extending the “good regulator theorem” - Nathaniel Virgo (PIBBSS Speaker Series) 2024

Towards an operationalisation of emergent phenomena - Fernando Rosas (PIBBSS Speaker Series) 2024

Jason Hoelscher-Obermaier - How LLM Evaluations Influence AI Risks

Nonequilibrium Neural Computation - Miguel Aguilera (PIBBSS Speaker Series) 2023

Cecilia Wood - Beyond vNM Self-modification and Reflective Stability - PIBBSS Symposium '23

Discovering the Adjacent Possible - Josh Bongard (PIBBSS Speaker Series) 2023

Nischal Mainali - A Geometry Viewpoint for Interpretability - PIBBSS Symposium '23

Matthew Lutz - Detecting emergent capabilities in multi-agent AI Systems - PIBBSS Symposium '23

Aysja Johnson - Searching For a Science of Abstraction - PIBBSS Symposium '23

Eleni Angelou - Overview of Problems in the Study of Language Model Behavior - PIBBSS Symposium '23

Gabriel Weil - Tort law as a tool for mitigating catastrophic risk from AI - PIBBSS Symposium '23

Guillaume Corlouer - The role of model degeneracy in the dynamics of SGD - PIBBSS Symposium '23

Ninell Oldenburg - Learning and Sustaining Social Norms as Normative Equilibria - Symposium '23

Sammy Martin - An overview of AI misuse risks and what to do about them - PIBBSS Symposium '23

Martin Soto - Constructing Logically Updateless Decision Theory - PIBBSS Symposium '23

Giles Howdle - Auto-Intentional Agency and AI Risk - PIBBSS Symposium '23

Nora Ammann - Navigating Challenges in Reasoning on Artificial and Intelligent Systems and Alignment

Potential value and trickiness of agency in biology - Alan Love (PIBBSS Speaker Series Chat) 2023

Challenges for Safe & Beneficial Brain-Like AGI - Steve Byrnes (PIBBSS Speaker Series) 2023

Paradigms of Alignment - Victoria Krakovna (PIBBSS Speaker Series) 2023

Bridging Gaps for AI Safety - Alan Chan (PIBBSS Speaker Series) 2023