INDEX
Explanations
mentions of different locations, cities, and cuisines
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.13
0.5%
1510
+0.12
0.4%
856
+0.10
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
227
+0.13
0.13
1199
+0.12
0.08
1510
+0.10
0.07
Negative Logits
<bos>
-3.71
Geplaatst
-1.04
kaarangay
-1.00
UnusedPrivate
-0.98
SourceChecksum
-0.97
Personendaten
-0.95
snippetHide
-0.94
nakalista
-0.93
Autoritní
-0.93
LookAnd
-0.90
POSITIVE LOGITS
kristal
0.90
kaos
0.87
Confu
0.86
optik
0.86
maske
0.85
lele
0.83
horrend
0.83
kase
0.83
stockholm
0.82
alike
0.81
Activations Density 1.939%