INDEX
Explanations
mentions of geek culture and related terms
references to geek culture
New Auto-Interp
Negative Logits
inished
-0.73
undai
-0.70
utions
-0.69
redistributed
-0.68
inating
-0.64
terminating
-0.64
Adin
-0.64
ugal
-0.64
comings
-0.63
inated
-0.61
POSITIVE LOGITS
geek
1.02
bench
0.99
hack
0.91
hattan
0.88
Dad
0.87
Crunch
0.86
nerd
0.84
haw
0.83
nerds
0.82
core
0.81
Activations Density 0.033%