INDEX
Explanations
mentions of the country Germany
references to Germany or German-related terms
New Auto-Interp
Negative Logits
ciating
-0.93
ctica
-0.84
efeated
-0.80
mble
-0.79
ilver
-0.77
vae
-0.77
heed
-0.76
okemon
-0.75
tis
-0.75
uably
-0.74
POSITIVE LOGITS
wings
1.19
Chancellor
1.02
shepherd
1.00
Bundesliga
0.85
Shepherd
0.83
oslov
0.82
Expression
0.81
Bundes
0.81
chancellor
0.77
Aerospace
0.75
Activations Density 0.024%