English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Adversarial examples can be useful for identifying vulnerabilities in AIsystems before they are deployed. In reinforcement learning (RL), adversarialpolicies can be developed by training an adversarial agent to minimize a targetagent's rewards. Prior work has studied black-box versions of these attackswhere the adversary only observes the world state and treats the target agentas any other part of the environment. However, this does not take into accountadditional structure in the problem. In this work, we study white-boxadversarial policies and show that having access to a target agent's internalstate can be useful for identifying its vulnerabilities. We make twocontributions. (1) We introduce white-box adversarial policies where anattacker observes both a target's internal state and the world state at eachtimestep. We formulate ways of using these policies to attack agents in2-player games and text-generating language models. (2) We demonstrate thatthese policies can achieve higher initial and asymptotic performance against atarget agent than black-box controls. Code is available athttps://github.com/thestephencasper/lm_white_box_attacks

Red Teaming with Mind Reading: White-Box Adversarial Policies Against RL Agents