INDEX
Explanations
words or phrases related to the Arabic language or culture
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
2
+0.15
0.9%
164
+0.11
0.6%
136
+0.11
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
503
+0.15
0.03
18
+0.11
0.03
417
+0.11
0.03
Negative Logits
CDATA
-1.58
blogger
-1.53
LED
-1.50
plit
-1.49
Argued
-1.41
unreadable
-1.38
built
-1.38
metab
-1.38
ignition
-1.37
xymatrix
-1.36
POSITIVE LOGITS
ytics
2.04
veolar
2.00
·
1.87
oupe
1.84
º
1.84
ī
1.83
ucci
1.81
cohol
1.80
Pradesh
1.76
ogue
1.73
Activations Density 0.187%