INDEX
Explanations
references to the publication "Forbes."
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1604
+0.17
0.9%
893
+0.14
0.8%
757
+0.14
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
227
+0.17
0.03
757
+0.14
0.02
893
+0.14
0.02
Negative Logits
<bos>
-1.20
Linki
-0.69
/**
-0.69
bibnamefont
-0.66
textTheme
-0.63
ngrx
-0.63
يتيمه
-0.61
Trotzdem
-0.59
Referències
-0.59
pinulongan
-0.59
POSITIVE LOGITS
Forbes
1.41
magis
1.32
Forbes
1.27
haup
1.25
ftre
1.24
fua
1.23
vns
1.22
kuro
1.21
jaya
1.19
Minang
1.18
Activations Density 0.185%