INDEX
Explanations
references to giveaways and contests
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
341
+0.10
0.3%
370
+0.08
0.2%
377
+0.07
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
377
+0.10
0.02
612
+0.08
0.02
1274
+0.07
0.02
Negative Logits
met
-0.71
doc
-0.70
Consulta
-0.68
met
-0.68
Documentos
-0.68
Doc
-0.67
prepare
-0.67
prepare
-0.66
vec
-0.66
/***
-0.65
POSITIVE LOGITS
giveaway
2.73
Giveaway
2.61
giveaways
2.54
affor
2.15
scrat
2.09
tupperware
2.07
swarovski
2.00
madonna
1.94
milf
1.94
maneu
1.93
Activations Density 0.176%