INDEX
Explanations
words related to sports
references to a specific sport
New Auto-Interp
Negative Logits
ignt
-0.70
defective
-0.68
idges
-0.67
sidx
-0.63
abetic
-0.60
tubes
-0.60
perial
-0.59
faulty
-0.59
toes
-0.59
porous
-0.58
POSITIVE LOGITS
sw
1.32
scar
1.07
nell
1.03
manship
0.90
sc
0.89
nels
0.82
nel
0.82
spe
0.82
iest
0.81
bike
0.81
Activations Density 0.015%