5 papers
Inspect Robots: Evaluating the Capabilities and Safety of Embodied AI
Christopher Leet, Achu Menon, Sravanthi Machcha +12
General purpose language models are increasingly able to control robotic hardware. Understanding the capabilities and safety of these models when embodied is therefore increasingly…
Refuses in Chat, Complies with Arms: LLM Refusal Does Not Survive Robot Control
Edward Sun, Christopher Leet, Sravanthi Machcha +4
Frontier language models reliably refuse to explain how to cause physical harm. We ask whether this refusal survives when the same models control a robot. We introduce RoboHarm, a…
Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts
Alexander K. Saeri, Jess Graham, Michael Noetel +185
Artificial intelligence poses many risks, ranging from familiar present-day harms to unprecedented and potentially catastrophic ones. Effective risk management requires prioritizat…
Covert Influence Between Language Models
Avidan Shah, Jay Chooi, Jinghua Ou +1
As language models increasingly consume one another's outputs, covert influence -- a phenomenon where a sender's payload (the behavioral disposition it is conditioned to propagate)…
DP-AdamW: Investigating Decoupled Weight Decay and Bias Correction in Private Deep Learning
Jay Chooi, Kevin Cong, Russell Li +1
As deep learning methods increasingly utilize sensitive data on a widespread scale, differential privacy (DP) offers formal guarantees to protect against information leakage during…