INDEX
Explanations
`case` followed by string literals
New Auto-Interp
Negative Logits
(
1.10
([
1.10
[
1.00
((
0.94
[
0.90
(
0.89
（
0.86
([
0.86
((-
0.86
[(
0.85
POSITIVE LOGITS
"
2.27
'
1.90
"_
1.88
"¿
1.87
`"
1.79
“
1.69
"!
1.69
"'
1.68
'_
1.67
="
1.66
Activations Density 0.836%