INDEX
Explanations
contrasting conjunctions that indicate exceptions or conditions
New Auto-Interp
Negative Logits
pite
-0.17
æĪIJ人
-0.15
hl
-0.15
IGHL
-0.14
inati
-0.14
çε
-0.14
reon
-0.14
Turns
-0.14
chner
-0.13
ocio
-0.13
POSITIVE LOGITS
otherwise
0.31
otherwise
0.28
åIJ¦
0.26
Otherwise
0.24
Otherwise
0.24
thus
0.23
OTHERWISE
0.21
thereby
0.19
thus
0.18
böylece
0.18
Activations Density 0.004%