Abstract
Training models to act in accordance with an explicitly defined set of principles, or "constitution," has shown promise as a robust and transparent mechanism for AI alignment. However, the generality and flexibility of such methods remain unclear. Here, we show that constitution-consistent behavior can be distilled from synthetic corpora into lightweight objects (low-rank adapters and steering vectors). Despite never seeing a harmful request or jailbreak during training, such objects increase jailbreak defense success and measured alignment -- particularly at long context lengths and against multi-turn attacks, where they outperform both prompted and steered baselines. Subtracting control-trained from constitution-trained objects further accentuates these effects, yielding defenses we call "constitutional adapters" (CAs). CAs can be trained on a base model, transferred zero-shot to its post-trained checkpoint, and scaled at inference time to predictably trade off defense for benign compliance. Taken together, these results recommend CAs as a lightweight, portable, and tunable lever for mitigating misalignment and misuse in API deployments.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Lowet, A. S., & Kurzeja, M. (2026). Constitutional adapters: Inference-time interventions for misalignment and misuse. https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse
MLA 9
Lowet, Adam S., and Mark Kurzeja. "Constitutional adapters: Inference-time interventions for misalignment and misuse." https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse.
Chicago (author–date)
Lowet, Adam S., and Mark Kurzeja. 2026. "Constitutional adapters: Inference-time interventions for misalignment and misuse." https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse.
Harvard
Lowet, A. S. and Kurzeja, M. (2026) 'Constitutional adapters: Inference-time interventions for misalignment and misuse', Available at: https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse.
Vancouver
Lowet AS, Kurzeja M. Constitutional adapters: Inference-time interventions for misalignment and misuse. https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse
IEEE
A. S. Lowet, and M. Kurzeja, "Constitutional adapters: Inference-time interventions for misalignment and misuse," https://omanscience.com/en/articles/constitutional-adapters-inference-time-interventions-for-misalignment-and-misuse.