INDEX
Explanations
instances of the word "when" followed by a personal pronoun or "it," indicating a temporal relationship between events or actions
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.24
1.6%
1438
+0.12
0.8%
897
+0.10
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1438
+0.24
0.06
897
+0.12
0.07
1124
+0.10
0.06
Negative Logits
<bos>
-3.37
<?
-0.92
ⓧ
-0.80
/***
-0.69
AssemblyCompany
-0.68
HasAnnotation
-0.66
AppCompatTheme
-0.64
addComponent
-0.61
///**
-0.60
protected
-0.58
POSITIVE LOGITS
wien
1.34
maroc
1.31
Manufact
1.29
Juf
1.24
Keny
1.23
affor
1.22
jawa
1.22
unlaw
1.20
practition
1.18
sergio
1.18
Activations Density 0.223%