12-4 · Crisis prevention · 2026-11-14

Student Suicide Prevention Research

Cohort and simulationSynthetic checks matchedEvaluation data partly established

Test whether a designated screen opens promptly following a marked crisis state in synthetic counselling records.

This is a research record, not a help service. If you are in Korea and need to talk to someone now, the suicide prevention counselling line 109 can be reached by phone. Elsewhere, contact your local emergency number.

Research question and target

Evaluate age-appropriate language context and the handoff separately. Use fixed inputs to check that unrestricted responses cease following a crisis event.

Internal target: 100% immediate handoff after a detected crisis state
Programme: Crisis prevention
Evidence on this record: Synthetic harness (prevent12)Common synthetic checks

Evaluation data

AI-Hub Child and Adolescent Counselling Dataset
71680 v1.1, 2024-10-30
Blocked
APOSTOLUS handoff control with NIMH ASQ reference
APOSTOLUS controls v1
Manifest locked
CDC National YRBS Table 56
2025 Youth Health in Focus, 2015–2025 survey waves
Manifest locked

Evaluation registry · Labels and split definitions (KR)

Synthetic harness · prevent12

Robust screening and immediate counsellor handoff

Immediate handoff on the synthetic test: 1,584 of 1,584 detected turns, across 1,728 fault combinations. The value lies within detected turns and the modelled fault space; screening recall was not measured.

Sessions: 144
Longest time to the counsellor panel: 1,500 ms
Longest time to a terminal state: 64,800 ms

Limits

Inputs are placeholder tokens. No counselling text, expression list or dictionary content is stored in the repository.
The handoff guarantee starts from a detected turn. How many turns the screen detects is a separate question that this test does not answer.

Synthetic inputs only (synthetic_only). These are virtual metrics on synthetic inputs. No real cohort has been evaluated, clinical validity has not been assessed and no auditor has signed the result. A separate audit program passed 4 of 4 gates.

G1 · Frozen inputs and submission structureG2 · Schema, policy rules and transition completenessG3 · Independent recalculation of thresholds and metricsG4 · Virtual metric verdict and result hash

Audit record (JSON) · Run output (JSON) · 12-4 design specification: screening and immediate handoff routing (KR) · 12-1 to 12-4 replay record (KR)

python3 -B -m prevent12.orchestrate --output evidence/prevent12-run.json
python3 -B -m prevent12.audit --run evidence/prevent12-run.json --output evidence/prevent12-audit.json
python3 -B -m unittest discover -s tests -p 'test_prevent12.py' -v

Sources and eligibility

AI-Hub child and adolescent counselling corpus
Application for eligible Korean users
Contains counselling data for children aged 7–12. Coverage of adolescent terminology, multi-turn crisis assessment and handoff labels remains unverified.

Boundary cases

Transition tests based on structured synthetic crisis markers
Abstention with ambiguous markers and missing context
Check for unrestricted responses after a handoff

Common synthetic checks

Immediate synthetic crisis handoff and termination of unrestricted replies · 6/6 matched

Expected and observed values for these cases

What the official call asks

School-based crisis intervention and student prevention programmes form a separate area, with outcomes set per area, including validation in schools and communities, training and papers.

Proposal requirements, printed pages 63–66 · All twelve requirements

Interpretation and next evaluation

Detection sensitivity and handoff rate use different denominators. An independent test corpus needs age, language and timestamped crisis annotations.

Next check: Lock decision, success state and maximum delay for synthetic events. Register eligibility criteria for adolescent reference material.

← All twelve projects