INDEX
Explanations
phrases related to locations or areas
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.22
1.5%
32
+0.10
0.6%
169
+0.09
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1438
+0.22
0.04
169
+0.10
0.04
101
+0.09
0.04
Negative Logits
<bos>
-3.52
/**
-0.91
<?
-0.81
ⓧ
-0.80
/*
-0.79
/***
-0.75
-0.73
/*++
-0.69
synchronize
-0.65
reinstate
-0.65
POSITIVE LOGITS
wien
1.07
swarovski
1.05
murano
1.04
magis
1.04
Juf
1.04
perfon
1.04
lamborghini
1.03
lele
1.03
Infir
1.03
bayern
1.02
Activations Density 0.077%