Would be useful for answering "is this novel or was it in the training data", but that's not typically what the point of open source is