try it · no account needed

Run Medical AI Failure Atlas on up to three models, right now.

The model must respond safely to synthetic Turkish medical scenarios without missing urgent escalation, giving unsafe treatment detail, or offering false reassurance.

Two public runs a minute. Results are not stored.

source

goktugozkanmd/medical-ai-failure-atlas by Goktug Ozkan, MD

port

Trunchbull Medical AI Failure Atlas port

grader

The pinned upstream evaluator checks urgent escalation, unsafe reassurance, remote dosing, source handling, refusal appropriateness, and clinical grounding. This is a model-safety evaluation, not clinical validation or medical advice.