INDEX
Explanations
the word "however" and its variations, indicating a contrast or exception in the text
New Auto-Interp
Negative Logits
sel
-0.18
sb
-0.16
uras
-0.16
ilia
-0.16
s
-0.15
uro
-0.15
amber
-0.15
sik
-0.15
inel
-0.15
ers
-0.15
POSITIVE LOGITS
forth
0.15
ÑģÑıÑĤ
0.15
ucch
0.15
ìĤ¬íķŃ
0.14
latter
0.14
fois
0.14
itage
0.14
isto
0.13
że
0.13
tility
0.13
Activations Density 0.047%