INDEX
Explanations
information related to a specific brand or product
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.14
0.5%
658
+0.13
0.5%
755
+0.10
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
658
+0.14
0.07
74
+0.13
0.04
1798
+0.10
0.06
Negative Logits
<bos>
-2.24
/*!
-0.66
collecte
-0.58
-0.58
referenties
-0.57
HasColumnType
-0.56
<!--
-0.56
Hå
-0.54
consultato
-0.53
дописавши
-0.53
POSITIVE LOGITS
eiffel
1.40
santiago
1.31
affor
1.31
swarovski
1.31
hornblende
1.29
nutella
1.27
peppa
1.27
leonardo
1.25
paradiso
1.24
embodi
1.24
Activations Density 0.681%