INDEX
Explanations
code snippets related to system configuration and setting up program options
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1699
+0.17
0.6%
876
+0.17
0.6%
304
+0.14
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
876
+0.17
0.00
304
+0.17
0.01
1622
+0.14
0.03
Negative Logits
s
-0.94
t
-0.94
c
-0.92
y
-0.92
m
-0.92
l
-0.91
f
-0.91
j
-0.90
p
-0.90
b
-0.89
POSITIVE LOGITS
stockholm
3.28
wien
3.19
embra
3.18
noel
3.18
milano
3.18
napoli
3.17
fta
3.08
frankfurt
3.05
squa
3.04
?...
3.04
Activations Density 0.199%