Xander Davies

Sparse Distributed Memory is a Continual Learner

D. Krotov, Gabriel Kreiman, Trenton Bricken, 2023, ICLR.

Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback

A. Dragan, Dmitrii Krasheninnikov, David Krueger, 2023, ArXiv.

Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback

A. Dragan, Dmitrii Krasheninnikov, David Krueger, 2023, Trans. Mach. Learn. Res..

Circuit Breaking: Removing Model Behaviors with Targeted Ablation

Xander Davies, Maximilian Li, Max Nadeau, 2023, ArXiv.

Discovering Variable Binding Circuitry with Desiderata

Tamar Rott Shaham, David Bau, N. Prakash, 2023, ArXiv.