INDEX
Explanations
references to various types of electronic devices and platforms as well as specific cultural events and personal life aspects
specific objects and contexts
New Auto-Interp
Negative Logits
informée
-0.92
ikusbot
-0.88
ſehen
-0.82
vooz
-0.82
ſchaft
-0.81
iſche
-0.81
erſten
-0.80
<unused41>
-0.79
<unused32>
-0.79
<unused8>
-0.79
POSITIVE LOGITS
.
0.33
ff
0.32
<eos>
0.29
,
0.28
↵
0.28
and
0.28
,
0.27
).
0.27
or
0.27
key
0.26
Activations Density 0.001%