INDEX
Explanations
instances where a specific term or name is mentioned
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
776
+0.15
0.5%
50
+0.15
0.5%
382
+0.12
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1984
+0.15
0.09
16
+0.15
0.09
1896
+0.12
0.05
Negative Logits
tramonto
-1.06
medesimo
-1.02
delà
-0.95
signore
-0.95
gardien
-0.89
créateur
-0.87
vainqueur
-0.86
mattino
-0.86
giudice
-0.85
silenzio
-0.84
POSITIVE LOGITS
Sklici
0.76
barran
0.68
Flere
0.67
dozen
0.67
Glej
0.67
Pä
0.66
তথ্যসূত্র
0.65
dozen
0.65
astéro
0.64
handful
0.64
Activations Density 0.515%