INDEX
Explanations
verbs related to actions or processes
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
460
+0.10
0.3%
559
+0.09
0.3%
1643
+0.08
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
460
+0.10
0.05
1350
+0.09
0.05
569
+0.08
0.05
Negative Logits
NameInMap
-0.59
stoff
-0.55
hek
-0.53
kundige
-0.52
Werden
-0.50
plak
-0.49
historie
-0.49
rech
-0.49
velg
-0.48
vermel
-0.48
POSITIVE LOGITS
esfuer
0.73
createDate
0.72
newVal
0.70
fileSize
0.70
fileList
0.67
venait
0.66
gomma
0.65
profondità
0.65
giù
0.65
loggedIn
0.65
Activations Density 0.338%