INDEX
Explanations
dollar amounts or budget-related information
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
382
+0.14
0.5%
381
+0.11
0.4%
1967
+0.10
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
658
+0.14
0.05
68
+0.11
0.04
382
+0.10
0.04
Negative Logits
delà
-1.10
Secrétaire
-1.07
levier
-1.04
Ministre
-1.02
Docteur
-1.00
bénéficiaire
-1.00
ruban
-0.99
désert
-0.97
exécu
-0.97
medesimo
-0.95
POSITIVE LOGITS
lot
0.83
tendency
0.72
great
0.69
huge
0.68
tremendous
0.66
chance
0.66
strong
0.65
plethora
0.63
good
0.61
few
0.61
Activations Density 0.190%