Jailbreak Scaling Laws for Large Language Models: Polynomial-Exponential Crossover
Quick summary
arXiv:2603.11331v4 Announce Type: replace-cross Abstract: Adversarial attacks can reliably steer safety-aligned large language models toward unsafe behavior. Empirically, we find that adversarial prompt-injection attacks can amplify attack success rate from the slow polynomial growth observed without injection to exponential growth with the number of inference-time samples. We first identify a minimal statistical mechanism for these two regimes by giving a small set of assumptions on the distribution of safe generation across contexts under which both scaling laws follow. To explain this pheno
Key takeaways
- arXiv:2603.11331v4 Announce Type: replace-cross Abstract: Adversarial attacks can reliably steer safety-aligned large language models toward unsafe behavior.
- Empirically, we find that adversarial prompt-injection attacks can amplify attack success rate from the slow polynomial growth observed without injection to exponential growth with the number of inference-time samples.
- We first identify a minimal statistical mechanism for these two regimes by giving a small set of assumptions on the distribution of safe generation across contexts under which both scaling laws follow.
Why it matters
“Jailbreak Scaling Laws for Large Language Models: Polynomial-Exponential Crossover” shows why AI risk cannot be reduced to answer accuracy. Access controls, logging, human approval and incident response need to be designed into the workflow from the start.
Member comments