No More, No Less: Least-Privilege Language Models

Paulius Rauba, Dominykas Seputis, Patrikas Vanagas, Mihaela Van Der Schaar
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:103980-104010, 2026.

Abstract

Least privilege is a core security principle: grant each request only the minimum access needed to achieve its goal. Deployed language models almost never follow it, instead being exposed through a single API endpoint that serves all users and requests. This gap exists not because least privilege would be unhelpful—deployments would benefit greatly from reducing unnecessary capability exposure. The real obstacle is definitional and mechanistic: what does "access" mean inside a language model, and how can we enforce it without retraining or deploying multiple models? We take inspiration from least privilege in computer systems and define a class of models called least-privilege language models, where privilege is reachable internal computation during the forward pass. In this view, lowering privilege literally shrinks the model’s accessible function class (as opposed to denying access via learned policies). We formalize deployment-time control as a monitor–allocator–enforcer stack, separating (i) request-time signals, (ii) a decision rule that allocates privilege, and (iii) an inference-time mechanism that selects privilege. We then propose Nested Least-Privilege Networks, a shape-preserving, rank-indexed intervention that provides a smooth, reversible control knob. We show that this knob yields policy-usable privilege–utility frontiers and enables selective suppression of targeted capabilities with limited collateral degradation across various policies. Most importantly, we see this as a defense of a completely new deployment paradigm which challenges the premise that we can only have output-level control of language models.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-rauba26a, title = {No More, No Less: Least-Privilege Language Models}, author = {Rauba, Paulius and Seputis, Dominykas and Vanagas, Patrikas and Van Der Schaar, Mihaela}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {103980--104010}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/rauba26a/rauba26a.pdf}, url = {https://proceedings.mlr.press/v306/rauba26a.html}, abstract = {Least privilege is a core security principle: grant each request only the minimum access needed to achieve its goal. Deployed language models almost never follow it, instead being exposed through a single API endpoint that serves all users and requests. This gap exists not because least privilege would be unhelpful—deployments would benefit greatly from reducing unnecessary capability exposure. The real obstacle is definitional and mechanistic: what does "access" mean inside a language model, and how can we enforce it without retraining or deploying multiple models? We take inspiration from least privilege in computer systems and define a class of models called least-privilege language models, where privilege is reachable internal computation during the forward pass. In this view, lowering privilege literally shrinks the model’s accessible function class (as opposed to denying access via learned policies). We formalize deployment-time control as a monitor–allocator–enforcer stack, separating (i) request-time signals, (ii) a decision rule that allocates privilege, and (iii) an inference-time mechanism that selects privilege. We then propose Nested Least-Privilege Networks, a shape-preserving, rank-indexed intervention that provides a smooth, reversible control knob. We show that this knob yields policy-usable privilege–utility frontiers and enables selective suppression of targeted capabilities with limited collateral degradation across various policies. Most importantly, we see this as a defense of a completely new deployment paradigm which challenges the premise that we can only have output-level control of language models.} }
Endnote
%0 Conference Paper %T No More, No Less: Least-Privilege Language Models %A Paulius Rauba %A Dominykas Seputis %A Patrikas Vanagas %A Mihaela Van Der Schaar %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-rauba26a %I PMLR %P 103980--104010 %U https://proceedings.mlr.press/v306/rauba26a.html %V 306 %X Least privilege is a core security principle: grant each request only the minimum access needed to achieve its goal. Deployed language models almost never follow it, instead being exposed through a single API endpoint that serves all users and requests. This gap exists not because least privilege would be unhelpful—deployments would benefit greatly from reducing unnecessary capability exposure. The real obstacle is definitional and mechanistic: what does "access" mean inside a language model, and how can we enforce it without retraining or deploying multiple models? We take inspiration from least privilege in computer systems and define a class of models called least-privilege language models, where privilege is reachable internal computation during the forward pass. In this view, lowering privilege literally shrinks the model’s accessible function class (as opposed to denying access via learned policies). We formalize deployment-time control as a monitor–allocator–enforcer stack, separating (i) request-time signals, (ii) a decision rule that allocates privilege, and (iii) an inference-time mechanism that selects privilege. We then propose Nested Least-Privilege Networks, a shape-preserving, rank-indexed intervention that provides a smooth, reversible control knob. We show that this knob yields policy-usable privilege–utility frontiers and enables selective suppression of targeted capabilities with limited collateral degradation across various policies. Most importantly, we see this as a defense of a completely new deployment paradigm which challenges the premise that we can only have output-level control of language models.
APA
Rauba, P., Seputis, D., Vanagas, P. & Van Der Schaar, M.. (2026). No More, No Less: Least-Privilege Language Models. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:103980-104010 Available from https://proceedings.mlr.press/v306/rauba26a.html.

Related Material