INDEX
Explanations
intensifiers, particularly the adverb "really."
New Auto-Interp
Negative Logits
aterno
-0.15
_EOL
-0.14
avery
-0.14
tres
-0.14
ificent
-0.14
érc
-0.14
'..
-0.14
urance
-0.14
ponse
-0.14
quite
-0.13
POSITIVE LOGITS
emann
0.17
enjoyed
0.17
aris
0.16
hit
0.16
luck
0.15
hope
0.15
uli
0.15
Mull
0.15
Kum
0.15
-bel
0.15
Activations Density 0.059%