INDEX
Explanations
expressions of fatigue or frustration
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.25
0.9%
1235
+0.09
0.3%
991
+0.08
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1476
+0.25
0.02
1235
+0.09
0.02
1852
+0.08
0.02
Negative Logits
<bos>
-2.68
/***
-0.68
ⓧ
-0.66
///**
-0.62
/**
-0.59
-0.58
require
-0.55
beforeAll
-0.54
Controllo
-0.52
autorytatywna
-0.52
POSITIVE LOGITS
riviera
1.13
toledo
1.09
frankfurt
1.06
tramont
1.03
verona
1.02
venice
1.01
Juf
0.99
eiffel
0.99
leonardo
0.99
ibiza
0.99
Activations Density 0.127%