Newest papers

Contextual Candour: Role-Aware Graduated Disclosure for LLMs over Mixed-Sensitivity Contexts

Devansh Srivastav, David Pape, Samira Abedini, Lea Schönherr · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28067 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Os LLMs operam cada vez mais em contextos de sensibilidade mista, onde devem raciocinar sobre informações que variam de público a confidencial, ao mesmo tempo em que revelam apenas o que o usuário está autorizado a receber. As abordagens existentes aplicam o controle de acesso como uma decisão binária: respondendo ou recusando totalmente, sem regulamentar o quanto ou que tipo de informação é revelada se o acesso for concedido, levando a um vazamento excessivo ou não intencional. Introduzimos a divulgação gradual, uma formulação de controle de acesso em que o nível de detalhes e sensibilidade na resposta de um modelo é calibrado para a autorização do usuário, em vez de ser tratado como uma decisão concedida ou negada. Devemos modelar a divulgação gradual como um problema de geração consciente de divulgação e propor

Machine translation (Meta M2M100); the English original follows.

Abstract

LLMs increasingly operate over mixed-sensitivity contexts, where they must reason over information ranging from public to confidential while disclosing only what the user is authorised to receive. Existing approaches enforce access control as a binary decision: either fully answering or refusing, without regulating how much or what kind of information is revealed if access is granted, leading to over-refusal or unintended leakage. We introduce graduated disclosure, an access-control formulation in which the level of detail and sensitivity in a model's response is calibrated to the user's authorisation rather than treated as a grant-or-deny decision. We model graduated disclosure as a disclosure-aware generation problem and propose a reward-based optimisation framework that shapes responses toward level-appropriate abstractions while penalising unauthorised factual leakage, including paraphrases and approximations. We evaluate our approach on enterprise and clinical benchmarks and compare it with prompt-based, gating, and supervised baselines. Experiments show that our method substantially reduces information leakage while preserving correctness and utility. These results suggest that disclosure control must be integrated into the generation process itself to enable secure yet useful LLM deployments.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Correctness, Confidentiality, Authorization, Access control, Computer science, Internet privacy

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange