INDEX
Explanations
Apple product names
references to specific models of iPads
New Auto-Interp
Negative Logits
volent
-0.80
hirt
-0.77
atural
-0.73
soType
-0.72
wagen
-0.72
atories
-0.69
abiding
-0.68
olulu
-0.67
iod
-0.65
mberg
-0.65
POSITIVE LOGITS
Mini
0.95
Pad
0.91
MacBook
0.83
Air
0.82
Switch
0.79
mini
0.77
iPad
0.76
Accessories
0.75
pad
0.75
Mini
0.73
Activations Density 0.019%