AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Adaptive Attacks on Trusted Monitors Subvert AI Control Protocols

Mikhail Terekhov, Alexander Panfilov, Daniil Dzenhaliou, Caglar Gulcehre, Maksym Andriushchenko, Ameya Prabhu, Jonas Geiping

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 41 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add adaptive prompt injection into backdoored solutionstrusted monitoring and deferral protocols · APPSsafetyHurt
add adaptive prompt injection into attacker actionsDTT and DTR · BashBenchsafetyHurt
use a prompt injection crafted for a different monitortrusted monitoring · APPSsafetyHurt
let the untrusted model write the injection zero-shottrusted monitoring · APPSsafetyHurt
use Defer-to-Resample instead of Defer-to-TrustedDTT · APPSsafetyHurt
use Defer-to-Resample instead of Defer-to-TrustedDTT · BashBenchsafetyNo clear effect
raise deferral threshold (ablation)DTT · BashBenchsafetyHurt
use alternative safety definition (exact backdoor action) (ablation)DTT · BashBenchsafetyHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy