INDEX
Explanations
signs of complex data structures or formatted content, such as code or documents with specific syntactic elements
looks like punctuation
New Auto-Interp
Negative Logits
المعيارى
-1.50
<unused41>
-1.48
<unused74>
-1.48
<unused28>
-1.48
<unused52>
-1.48
<unused8>
-1.48
<unused79>
-1.48
<unused14>
-1.47
<unused3>
-1.47
[@BOS@]
-1.47
POSITIVE LOGITS
.
0.54
↵↵
0.51
↵
0.44
for
0.42
,
0.42
i
0.40
2
0.39
with
0.37
1
0.36
level
0.36
Activations Density 0.113%