Skip to main content
Search

# Evolution Under Fire

1 related articles

Subscribe to updates
  1. arXiv

    Safety harnesses stop being hand-rewritten: Evolution Under Fire evolves attacks and defenses together, cutting injection success to 3.0%

    Tool-using language-model agents read emails, webpages and documents, so malicious instructions and legitimate task directives arrive through the same channel, while existing defenses are largely hand-designed and static. Evolution Under Fire couples evolution of complete executable safety harnesses with evolution of adaptive stressors in three stages: a utility-constrained Pareto tree search rewrites the harness, utility-preserving injections are generated, and those stressors feed back into further harness search. On AgentDyn with GPT-4o-mini as the task model, the evolved harness reduces attack success from 50.4% to 3.0% while raising clean and attacked utility.

You’re all caught up