INDEX
Explanations
words related to virtual data storage and network connections
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1978
+0.18
0.6%
297
+0.14
0.5%
1265
+0.11
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1145
+0.18
0.07
1142
+0.14
0.05
297
+0.11
0.07
Negative Logits
<bos>
-1.40
película
-1.06
Семья
-0.92
Galería
-0.87
Perú
-0.87
Iglesia
-0.86
Από
-0.85
Пото
-0.84
spesies
-0.83
Eksteraj
-0.82
POSITIVE LOGITS
aen
2.41
increa
2.40
mef
2.38
inev
2.37
fta
2.34
wien
2.32
effe
2.31
squa
2.25
desir
2.23
fuf
2.21
Activations Density 0.250%