INDEX
Explanations
references to specific geographic locations
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
156
+0.20
1.2%
376
+0.13
0.7%
74
+0.11
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
180
+0.20
0.01
74
+0.13
0.01
259
+0.11
0.01
Negative Logits
ĥ½
-3.72
¼
-3.38
ĺ
-3.34
į
-3.09
Į
-3.08
Ĵ
-3.00
¸
-2.94
İ
-2.91
ı
-2.84
·
-2.80
POSITIVE LOGITS
ue
2.04
dale
2.03
bank
2.00
field
1.83
light
1.79
fan
1.77
fields
1.69
Bank
1.65
iero
1.56
bourne
1.56
Activations Density 0.018%