INDEX
Explanations
mentions of the boxer Muhammad Ali
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
390
+0.18
0.8%
489
+0.14
0.6%
1034
+0.13
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
390
+0.18
0.03
1034
+0.14
0.02
1335
+0.13
0.02
Negative Logits
Lakeland
-0.51
hyperplasia
-0.50
thargy
-0.46
CascadeType
-0.46
earnestness
-0.45
lethargy
-0.44
LETIN
-0.43
ausgeschlossen
-0.43
INESS
-0.43
atience
-0.42
POSITIVE LOGITS
Ali
1.48
Ali
1.42
ali
1.24
ALI
1.08
Alias
0.94
minimalis
0.87
aliases
0.86
alias
0.86
Alias
0.86
Alli
0.84
Activations Density 0.088%