107 citations · 311 across the 8 of their papers we have counts for
8 papers
The Hydra Effect: Emergent Self-repair in Language Model Computations
Thomas McGrath, Matthew Rahtz, Janos Kramar +2
We investigate the internal structure of language model computations using causal analysis and demonstrate two motifs: (1) a form of adaptive computation where ablations of one att…
Power-seeking can be probable and predictive for trained agents
Victoria Krakovna, Janos Kramar
Power-seeking behavior is a key source of risk from advanced AI, but our theoretical understanding of this phenomenon is relatively limited. Building on existing theoretical result…
OpenSpiel: A Framework for Reinforcement Learning in Games
Marc Lanctot, Edward Lockhart, Jean-Baptiste Lespiau +24
OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games. OpenSpiel supports n-player (single- and multi…
Learning Reciprocity in Complex Sequential Social Dilemmas
Tom Eccles, Edward Hughes, János Kramár +2
Reciprocity is an important feature of human social interaction and underpins our cooperative nature. What is more, simple forms of reciprocity have proved remarkably resilient in…
Relational Forward Models for Multi-Agent Learning
Andrea Tacchetti, H. Francis Song, Pedro A. M. Mediano +5
The behavioral dynamics of multi-agent systems have a rich and orderly structure, which can be leveraged to understand these systems, and to improve how artificial agents learn to…
Reinforcement and Imitation Learning for Diverse Visuomotor Skills
Yuke Zhu, Ziyu Wang, Josh Merel +8
We propose a model-free deep reinforcement learning method that leverages a small amount of demonstration data to assist a reinforcement learning agent. We apply this approach to r…
Guidelines for Artificial Intelligence Containment
James Babcock, Janos Kramar, Roman V. Yampolskiy
With almost daily improvements in capabilities of artificial intelligence it is more important than ever to develop safety software for use by the AI research community. Building o…
The AGI Containment Problem
James Babcock, Janos Kramar, Roman Yampolskiy
There is considerable uncertainty about what properties, capabilities and motivations future AGIs will have. In some plausible scenarios, AGIs may pose security risks arising from…