INDEX
Explanations
words related to education and alumni
references to alumni and their achievements
New Auto-Interp
Negative Logits
plays
-0.75
erness
-0.69
lihood
-0.67
eah
-0.65
baby
-0.64
roid
-0.62
frantic
-0.61
Mart
-0.61
hamm
-0.61
omsday
-0.60
POSITIVE LOGITS
umni
1.35
alumni
1.32
uates
1.25
alum
1.23
graduates
0.95
umn
0.83
uating
0.79
uate
0.79
coh
0.78
uated
0.76
Activations Density 0.007%