INDEX
Explanations
numerical quantities followed by a unit of measure
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1385
+0.15
0.5%
994
+0.13
0.4%
394
+0.11
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
994
+0.15
0.04
776
+0.13
0.07
1526
+0.11
0.05
Negative Logits
Tikang
-0.73
Paglinawan
-0.71
rekon
-0.69
RTLR
-0.65
Ceinture
-0.63
IUrlHelper
-0.61
المناصب
-0.61
сылкі
-0.60
ujednoznacz
-0.60
katastro
-0.60
POSITIVE LOGITS
impra
1.14
maneu
1.05
indestru
1.01
reluct
1.00
unspeak
0.98
hentai
0.97
intersper
0.97
apprehen
0.96
Shakspeare
0.95
withal
0.95
Activations Density 0.649%