Complete result settings for Reuters-52 dataset (table columns are sortable)
The experiments are described in Rooshenas and Lowd, Discriminative Structure Learning of Arithmetic Circuits, AIStats 16

Standard Deviation	L1 Penalty	Train CLL	Validation CLL	Test CLL	Node#	Edge#	Feature#	Learning Time
0.1	0.1	-44.808872	-43.865229	-42.515507	838578	1916823	8232	57937.112219
0.5	0.1	-44.763277	-43.845101	-42.501417	838578	1916823	8232	62472.261772
0.5	2	-44.801087	-43.710675	-42.378747	838578	1916823	8232	77503.830626
0.1	2	-44.803971	-43.718853	-42.381535	838578	1916823	8232	84981.251885