INDEX
Explanations
names of locations or places, like cities and towns
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
251
+0.12
0.5%
395
+0.10
0.4%
1044
+0.09
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
648
+0.12
0.05
227
+0.10
0.06
1013
+0.09
0.06
Negative Logits
<bos>
-1.97
ManyToMany
-0.66
Шаг
-0.59
вающий
-0.59
خصة
-0.58
אַ
-0.58
Показать
-0.56
HasIndex
-0.56
Пото
-0.56
אַ
-0.56
POSITIVE LOGITS
thut
1.73
aen
1.49
fta
1.43
increa
1.41
mef
1.41
depic
1.38
fup
1.38
reft
1.38
madonna
1.38
ohr
1.36
Activations Density 0.440%