BETA-MIND Research Instrument
An illustrative prototype of the BETA-MIND reflection-instrument design: pick one of four human-agency dimensions, write a short reflection, and get a Socratic follow-up — never a score or a diagnosis.
Not a validated assessment — an illustrative prototype of the instrument's design
BETA-MIND (see the related project record) is a proposal-ready research instrument design; it has not been empirically validated. This demo never produces a score, rating, or diagnosis of any kind, and the agent is not a therapist or clinician. It responds with a brief, non-judgmental reflection and one open follow-up question. Your reflection text is sent to the model to generate a reply and is not stored anywhere afterward. If your reflection suggests you may be in crisis, the agent will pause the exercise and point you to real help — in the US, call or text 988 (Suicide & Crisis Lifeline) any time, free and confidential.
How it works
Pick one of the four BETA-MIND dimensions — independent reasoning, socialization, creativity, or decision accountability — read its reflection prompt, and write a few sentences about a real or recent moment. The reflection guide agent reflects back what you described and asks one open question to sit with, in the same Socratic spirit as the instrument design described in the BETA-MIND project record.
Try it
Select a dimension to see its reflection prompt.
Loading dimensions…
If you're going through something difficult, please reach out to a mental health professional or a crisis line — in the US, call or text 988 any time.
Why this exists
This is a direct, hands-on companion to the BETA-MIND project record: rather than only describe the four agency dimensions on paper, this lets a visitor actually try the kind of reflective prompt the instrument proposes, and see the deliberately non-scoring, non-diagnostic response style the design calls for.
Production checklist
- No numeric score, rating, or diagnosis is ever produced — by design, not just by convention.
- The agent is explicitly instructed it is not a therapist or clinician.
- A crisis-resource message replaces the reply whenever content is flagged as high-risk.
- Reflection text is sent only to the model call for this one request; nothing is persisted.
- Automated tests cover a normal reflection and the crisis-safety fallback path.