INDEX
Explanations
proper names, specifically related to characters and personalities
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
629
+0.13
0.6%
812
+0.12
0.5%
966
+0.11
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
629
+0.13
0.02
1056
+0.12
0.03
966
+0.11
0.02
Negative Logits
дивиду
-0.47
VolleyError
-0.47
OGND
-0.46
viaf
-0.44
пись
-0.43
late
-0.43
</em>
-0.43
IBRATION
-0.43
죽
-0.42
лоси
-0.41
POSITIVE LOGITS
FRANK
1.49
FRANK
1.40
Frank
1.40
Frank
1.38
frank
1.24
Franks
1.22
frank
1.06
fran
1.02
FRANKLIN
1.00
Franklin
0.93
Activations Density 0.092%