INDEX
Explanations
connective words and phrases signaling relationships between different parts of a text
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1445
+0.16
0.5%
314
+0.11
0.4%
1892
+0.11
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1921
+0.16
0.07
1892
+0.11
0.06
1806
+0.11
0.06
Negative Logits
antik
-1.23
soggior
-1.15
teras
-1.13
keramik
-1.12
Kategor
-1.11
alkoh
-1.09
exé
-1.09
meras
-1.09
elek
-1.08
makro
-1.08
POSITIVE LOGITS
whatnot
1.00
therefor
0.93
intersper
0.92
prolly
0.91
they
0.86
unavoid
0.86
we
0.84
detest
0.81
miscon
0.81
vainly
0.81
Activations Density 0.374%