Table 6. Ensemble-classifiers.

The size of the tf-idf vectors in the tf-idf baseline is n = 100k.

Classifier	Training- and test data
	s		xl		s-k		xl-k		GPT3		Grover
	Acc.	AUC	Acc.	AUC	Acc.	AUC	Acc.	AUC	Acc.	AUC	Acc.	AUC
Baselines
Feature-baseline	0.897	0.964	0.759	0.836	0.927	0.975	0.858	0.932	0.779	0.859	0.692	0.767
tf-idf-baseline	0.855	0.935	0.710	0.787	0.959	0.993	0.915	0.972	0.749	0.837	0.690	0.764
Ensembles
LR sep.	0.877	0.959	0.740	0.831	0.966	0.995	0.920	0.976	0.761	0.844	0.689	0.764
NN sep.	0.918	0.973	0.782	0.877	0.971	0.995	0.924	0.975	0.786	0.862	0.724	0.804
LR super	0.880	0.957	0.714	0.802	0.962	0.991	0.912	0.969	0.754	0.853	0.691	0.783
NN super	0.882	0.957	0.716	0.803	0.961	0.988	0.905	0.965	0.774	0.864	0.716	0.805