INDEX
Explanations
words related to or containing the string "irk"
instances of the word "quirk" and its variations
New Auto-Interp
Negative Logits
ccording
-0.81
++++
-0.72
eleph
-0.66
Þ
-0.66
receptive
-0.66
milo
-0.65
president
-0.61
tremend
-0.61
srf
-0.61
·
-0.60
POSITIVE LOGITS
irk
1.21
patrick
1.15
weed
0.86
lich
0.83
itives
0.82
nesses
0.82
bys
0.78
ishly
0.78
ness
0.78
wall
0.76
Activations Density 0.009%