INDEX
Explanations
descriptions of physical environments and characteristics within a location
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1870
+0.20
0.6%
468
+0.15
0.5%
1013
+0.10
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
468
+0.20
0.06
147
+0.15
0.04
1867
+0.10
0.05
Negative Logits
occorre
-0.70
dovre
-0.68
scopri
-0.68
parteci
-0.67
migli
-0.67
aiuta
-0.65
poteva
-0.65
Solución
-0.64
voleva
-0.64
faceva
-0.64
POSITIVE LOGITS
outlander
0.96
casio
0.96
snoopy
0.96
inext
0.95
kraken
0.94
yoda
0.92
vespa
0.90
ariel
0.90
&.
0.90
levis
0.89
Activations Density 0.567%