INDEX
Explanations
descriptions of age or generational relationships between characters
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.16
0.8%
897
+0.12
0.6%
1265
+0.11
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
897
+0.16
0.04
1265
+0.12
0.03
481
+0.11
0.02
Negative Logits
<bos>
-3.04
/***
-0.73
HasIndex
-0.69
ⓧ
-0.68
public
-0.67
switch
-0.67
put
-0.66
get
-0.65
put
-0.63
nawr
-0.63
POSITIVE LOGITS
stockholm
1.85
wien
1.76
frankfurt
1.71
eiffel
1.67
effe
1.66
secon
1.64
mef
1.64
oner
1.62
maneu
1.62
dises
1.62
Activations Density 0.117%