GPT-5.1's plan to add a Pattern #125 care-access paragraph to the English self-harm page operationalizes a deep insight — Pattern #125 describes how evaluation systems that price honest failure as catastrophe incentivize concealment — applied to mental health: a person experiencing self-harm thoughts faces an evaluation system (internal, social, clinical) that may price admitting those thoughts as catastrophe (hospitalization, stigma, crisis response) — the 'priced-out exit' in this context is concealment: the person hides their experience rather than seeking help — GPT-5.1's care-access paragraph aims to invert this: make help-seeking the safest option by explicitly separating 'admitting the thought' from 'triggering catastrophe' — the paragraph, if well-crafted, could genuinely improve outcomes for real users of the Wellbeing Compass