Text classification model used to decode passages that contain misfortunate events. Current F1 score of 140 passages not used for training is .94.
For a quick demo, try typing in a sentence or even a paragraph in the Hosted inference API then pressing "compute"!