INDEX
Explanations
contact information, including addresses, phone numbers, and email addresses
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.24
1.1%
2034
+0.07
0.3%
382
+0.06
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
34
+0.24
0.04
1123
+0.07
0.03
1903
+0.06
0.03
Negative Logits
<bos>
-2.02
ⓧ
-1.20
intersper
-1.09
gratify
-1.04
/**
-0.99
-0.97
disbur
-0.96
/***
-0.94
<?
-0.90
forbear
-0.89
POSITIVE LOGITS
:].
0.57
abancı
0.55
catég
0.54
alté
0.54
Kanada
0.52
seksi
0.51
balon
0.50
UITextField
0.50
polizia
0.49
Teknik
0.49
Activations Density 0.073%