INDEX
Explanations
conditional statements within programming code, and potentially temporal references in regular text.
New Auto-Interp
Negative Logits
purpoſe
-1.37
ſelf
-1.37
Majefty
-1.36
themſelves
-1.34
^(@)
-1.34
Efq
-1.33
whoſe
-1.31
myſelf
-1.29
becauſe
-1.29
himſelf
-1.23
POSITIVE LOGITS
<bos>
1.38
↵
1.30
↵↵
1.11
"
0.98
e
0.95
a
0.94
'
0.93
<eos>
0.90
i
0.90
n
0.90
Activations Density 2.746%