INDEX
Explanations
references to fairies and related mythical figures or themes
New Auto-Interp
Negative Logits
ersh
-0.16
orious
-0.15
ips
-0.15
ummies
-0.15
asse
-0.15
erc
-0.14
oner
-0.14
ating
-0.14
idges
-0.14
mons
-0.14
POSITIVE LOGITS
tale
0.29
Tale
0.28
dust
0.26
land
0.25
tales
0.24
Dust
0.22
dust
0.22
ta
0.21
god
0.20
Tales
0.18
Activations Density 0.006%