INDEX
Explanations
descriptions related to objects, construction, and quality
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1403
+0.17
0.6%
1385
+0.17
0.5%
297
+0.14
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1403
+0.17
0.02
185
+0.17
0.04
1446
+0.14
0.03
Negative Logits
nguyen
-0.79
nikah
-0.79
passim
-0.77
Muhamma
-0.77
Öster
-0.76
Tadeusz
-0.75
Nguy
-0.75
Rumania
-0.74
pamph
-0.73
Wakil
-0.73
POSITIVE LOGITS
compagnon
0.97
époux
0.94
rassemble
0.84
swarovski
0.83
automne
0.83
célé
0.81
Février
0.81
vété
0.80
Décembre
0.80
Australie
0.80
Activations Density 0.505%