Silent Manipulation of Mental Health Treatment Recommendations from a Large Language Model

Abstract excerpt

<h4>Importance</h4> Large language models (LLMs) increasingly inform mental health decisions by patients and clinicians. Inference-time activation steering can shift model behavior on a target dimension without altering weights or prompts and without disclosure to users, allowing treatment recommendations to be silently changed for commercial or ideological reasons. <h4>Objective</h4> To determine whether direct...

Topics

Open a Topic to create a Post that cites this publication.

• Antidepressive Agents
• Australia
• Genetic Risk Score
• Genome-Wide Association Study
• Language
• Major Depressive Disorder
• Phenotype
• Selective Serotonin Reuptake Inhibitors

Identifiers and source

Literature Corpus work

4740bc35-a782-5ce7-bcd5-e0f78ec38458

DOI

10.64898/2026.06.16.26355686