INDEX
Explanations
numerical sequences containing increasing or decreasing values
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
964
+0.17
0.5%
453
+0.13
0.4%
1978
+0.13
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
964
+0.17
0.05
227
+0.13
0.05
981
+0.13
0.05
Negative Logits
wieś
-0.88
surpl
-0.88
étu
-0.83
préc
-0.80
soigne
-0.75
renfer
-0.74
mène
-0.70
prouve
-0.69
paraît
-0.69
réuss
-0.68
POSITIVE LOGITS
keramik
1.30
alkoh
1.27
optik
1.26
akut
1.22
karton
1.20
antik
1.15
kosme
1.14
panik
1.14
kooper
1.14
kram
1.14
Activations Density 0.229%