INDEX
Explanations
phrases related to governmental activities and consultations
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
678
+0.11
0.3%
1499
+0.10
0.3%
80
+0.07
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
678
+0.11
0.05
1207
+0.10
0.04
968
+0.07
0.05
Negative Logits
swarovski
-1.35
bordeaux
-1.33
ecru
-1.30
!...
-1.26
Mlle
-1.25
madonna
-1.24
emphat
-1.21
ftu
-1.20
guarante
-1.19
embra
-1.15
POSITIVE LOGITS
ways
0.69
strategies
0.65
possible
0.65
solutions
0.64
ideas
0.62
possibilities
0.62
how
0.61
issues
0.61
脚注の使い方
0.61
recommendations
0.60
Activations Density 0.463%