INDEX
Explanations
mentions of Pittsburgh and its sports teams
New Auto-Interp
Negative Logits
ope
-0.17
ergus
-0.17
uble
-0.16
uchar
-0.15
uddenly
-0.14
itas
-0.14
udas
-0.14
eya
-0.14
asaki
-0.14
ushima
-0.14
POSITIVE LOGITS
tails
0.17
ro
0.15
dar
0.15
riott
0.14
zan
0.14
asic
0.14
ÑĢон
0.14
argument
0.14
Tart
0.14
tile
0.13
Activations Density 0.005%