Complete result settings for MSNBC 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	0.5	-1.544084	-1.548626	-1.559523	213127	431453	10034	5889.945591
0.5	0.5	-1.544084	-1.548626	-1.559523	213127	431453	10034	8058.626902
0.5	1	-1.544085	-1.548627	-1.559522	213127	431453	10034	8886.637025
0.1	2	-1.544086	-1.548627	-1.559521	213127	431453	10034	9531.093053
0.5	2	-1.544086	-1.548627	-1.559521	213127	431453	10034	9189.547976
0.1	1	-1.544085	-1.548627	-1.559522	213127	431453	10034	11632.259627
0.5	0.1	-1.544083	-1.548625	-1.559521	213127	431453	10034	9017.567120
0.1	0.1	-1.544083	-1.548625	-1.559521	213127	431453	10034	9809.807682