INDEX
Explanations
mentions of percentages and statistics
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1577
+0.23
0.9%
394
+0.21
0.8%
1150
+0.19
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
394
+0.23
0.11
1314
+0.21
0.13
1577
+0.19
0.17
Negative Logits
<bos>
-0.89
GraphicsUnit
-0.79
Portály
-0.65
expandindo
-0.62
bewerken
-0.62
Normdatei
-0.62
abestanden
-0.61
виправивши
-0.60
లాలు
-0.59
aarrggbb
-0.59
POSITIVE LOGITS
impra
0.53
baju
0.52
fendi
0.52
₁(
0.51
uye
0.51
₁,
0.51
Minang
0.50
Kini
0.50
lada
0.50
puto
0.49
Activations Density 5.508%