INDEX
Explanations
references to widespread issues or actions
instances of the term "widespread."
New Auto-Interp
Negative Logits
ek
-0.73
ude
-0.70
craft
-0.67
Alone
-0.65
fuck
-0.62
ef
-0.62
curves
-0.62
gu
-0.61
Looks
-0.61
Presents
-0.61
POSITIVE LOGITS
widespread
3.33
pervasive
2.08
idespread
2.06
rampant
1.76
ubiquitous
1.68
prevalent
1.60
widely
1.52
endemic
1.45
commonplace
1.45
ubiqu
1.34
Activations Density 0.021%