INDEX
Explanations
phrases related to overcoming obstacles or finding workarounds
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.25
1.3%
61
+0.11
0.5%
1187
+0.10
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1187
+0.25
0.04
61
+0.11
0.04
662
+0.10
0.04
Negative Logits
<bos>
-2.70
//};
-0.75
ⓧ
-0.73
-0.67
/***
-0.65
HasIndex
-0.65
contentLoaded
-0.64
Argumento
-0.62
},{
-0.62
/**
-0.60
POSITIVE LOGITS
maroc
1.17
accla
1.15
maneu
1.15
affor
1.13
bandung
1.09
suspic
1.07
Juf
1.06
conflic
1.04
Batam
1.04
véhic
1.02
Activations Density 0.213%