INDEX
Explanations
references to the Marine Corps
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.17
1.0%
1124
+0.09
0.6%
920
+0.09
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
920
+0.17
0.02
517
+0.09
0.02
144
+0.09
0.02
Negative Logits
<bos>
-2.85
-0.69
/**
-0.68
introduce
-0.64
gather
-0.63
/*
-0.62
<?
-0.62
ⓧ
-0.61
evacuate
-0.61
prepare
-0.61
POSITIVE LOGITS
Minang
1.32
Juf
1.27
Bibl
1.22
véhic
1.21
Meksi
1.21
Jambi
1.20
Marine
1.19
soulign
1.17
hcm
1.16
tramont
1.15
Activations Density 0.081%