INDEX
Explanations
words related to snow or winter
references to snow
New Auto-Interp
Negative Logits
ENTION
-0.84
Jinn
-0.73
aird
-0.71
plur
-0.71
ysis
-0.70
###
-0.67
########
-0.66
ISTER
-0.64
entric
-0.64
ister
-0.63
POSITIVE LOGITS
flake
1.70
fall
1.14
Leopard
1.01
balls
1.00
storm
0.99
storms
0.97
boarding
0.97
tro
0.96
mobile
0.94
falls
0.92
Activations Density 0.022%