Effects of Scale on Language Model Robustness

no code implementations • 25 Jul 2024 • Nikolaus Howe, Ian McKenzie, Oskar Hollinsworth, Michał Zajac, Tom Tseng, Aaron Tucker, Pierre-Luc Bacon, Adam Gleave

Even with the advantage conferred by scale, undefended models remain easy to attack in absolute terms, and we thus turn our attention to explicitly training models for adversarial robustness, which we show to be a much more compute-efficient defense than scaling model size alone.

Adversarial Robustness Language Modeling +1

Can Go AIs be adversarially robust?

no code implementations • 18 Jun 2024 • Tom Tseng, Euan McLean, Kellin Pelrine, Tony T. Wang, Adam Gleave

Prior work found that superhuman Go AIs can be defeated by simple adversarial strategies, especially "cyclic" attacks.

Diversity

Inverse Scaling: When Bigger Isn't Better

no code implementations • 15 Jun 2023 • Ian R. McKenzie, Alexander Lyzhov, Michael Pieler, Alicia Parrish, Aaron Mueller, Ameya Prabhu, Euan McLean, Aaron Kirtland, Alexis Ross, Alisa Liu, Andrew Gritsevskiy, Daniel Wurgaft, Derik Kauffman, Gabriel Recchia, Jiacheng Liu, Joe Cavanagh, Max Weiss, Sicong Huang, The Floating Droid, Tom Tseng, Tomasz Korbak, Xudong Shen, Yuhui Zhang, Zhengping Zhou, Najoung Kim, Samuel R. Bowman, Ethan Perez

Here, we present evidence for the claim that LMs may show inverse scaling, or worse task performance with increased scale, e. g., due to flaws in the training objective and data.

Adversarial Policies Beat Superhuman Go AIs

2 code implementations • 1 Nov 2022 • Tony T. Wang, Adam Gleave, Tom Tseng, Kellin Pelrine, Nora Belrose, Joseph Miller, Michael D. Dennis, Yawen Duan, Viktor Pogrebniak, Sergey Levine, Stuart Russell

The core vulnerability uncovered by our attack persists even in KataGo agents adversarially trained to defend against our attack.