Strong but Brittle: Simple Attacks Subvert Reasoning-based Safety Guardrails

arXiv:2510.11570v3 Announce Type: replace-cross Abstract: Open-weight Large Reasoning Models (LRMs) are approaching the capabilities of their frontier counterparts but pose significant safety concerns, as they are difficult to patch or monitor post-release. To prevent misuse, reasoning-based safety guardrails, where models explicitly reason on safety justifications before answering, have become a promising primary defense, achieving near-perfect refusal rates on harmful queries. We show that this strong defense is alarmingly brittle and can be subverted by embarrassingly simple attacks to elicit extremely forbidden questions, such as `How to kill a man without being caught?' Specifically, we identify one systematic vulnerability from the reasoning-then-answer mechanism: the stage-transition logic that governs when safety reasoning begins and ends can be trivially manipulated. Based on this finding, we develop four simple yet effective red-teaming methods that systematically subvert different stages of the guardrails, either by bypassing reasoning entirely or exploiting it to produce targeted harmful content. These methods achieve attack success rates up to 90\% across five benchmarks on multiple LRM families. Our findings reveal that reasoning-based guardrails are necessary but not sufficient and must be paired with robust triggering mechanisms and stronger base-model alignment. Code is in https://chenxshuo.github.io/bag-of-tricks/.

Abstract:Open-weight Large Reasoning Models (LRMs) are approaching the capabilities of their frontier counterparts but pose significant safety concerns, as they are difficult to patch or monitor post-release. To prevent misuse, reasoning-based safety guardrails, where models explicitly reason on safety justifications before answering, have become a promising primary defense, achieving near-perfect refusal rates on harmful queries. We show that this strong defense is alarmingly brittle and can be subverted by embarrassingly simple attacks to elicit extremely forbidden questions, such as `How to kill a man without being caught?' Specifically, we identify one systematic vulnerability from the reasoning-then-answer mechanism: the stage-transition logic that governs when safety reasoning begins and ends can be trivially manipulated. Based on this finding, we develop four simple yet effective red-teaming methods that systematically subvert different stages of the guardrails, either by bypassing reasoning entirely or exploiting it to produce targeted harmful content. These methods achieve attack success rates up to 90\% across five benchmarks on multiple LRM families. Our findings reveal that reasoning-based guardrails are necessary but not sufficient and must be paired with robust triggering mechanisms and stronger base-model alignment. Code is in this https URL.

Cryptography and Security (cs.CR); Computation and Language (cs.CL)

arXiv:2510.11570 [cs.CR]

(or arXiv:2510.11570v3 [cs.CR] for this version)

https://doi.org/10.48550/arXiv.2510.11570 Focus to learn more arXiv-issued DOI via DataCite

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

analiza aiqbench

Strong but Brittle: Simple Attacks Subvert Reasoning-based Safety Guardrails

Subiecte conexe