INDEX
Explanations
references to temperature or heat
New Auto-Interp
Negative Logits
featureID
-0.79
DockStyle
-0.78
multer
-0.71
commissioning
-0.69
purpoſe
-0.68
itſelf
-0.66
werd
-0.63
Galer
-0.63
himſelf
-0.63
fubject
-0.62
POSITIVE LOGITS
preferred
0.77
Hot
0.74
prefer
0.72
Preferences
0.71
Hot
0.70
PREFERRED
0.68
hottest
0.67
Preferred
0.66
préféré
0.64
preferable
0.63
Activations Density 0.150%