INDEX
Explanations
references to burial and related actions
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.20
1.0%
82
+0.13
0.6%
1604
+0.12
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1013
+0.20
0.03
1437
+0.13
0.02
1604
+0.12
0.02
Negative Logits
<bos>
-2.07
cout
-0.65
/**
-0.62
maxcdn
-0.62
ids
-0.61
Vegeu
-0.61
quad
-0.61
Pró
-0.60
const
-0.60
times
-0.60
POSITIVE LOGITS
maneu
1.58
ecru
1.56
inconce
1.53
broderie
1.51
impra
1.48
snoopy
1.47
indestru
1.46
bordeaux
1.45
hentai
1.44
newArr
1.43
Activations Density 0.093%