INDEX
Explanations
formal closings and signatures
New Auto-Interp
Negative Logits
でも
1.03
ignores
0.97
weird
0.95
ignore
0.92
bizarre
0.91
interessant
0.89
strange
0.89
absurd
0.88
useless
0.87
그래도
0.86
POSITIVE LOGITS
Signature
0.78
Signature
0.72
/[
0.70
([
0.68
Sincerely
0.68
[
0.66
signature
0.66
0.63
0.63
________________
0.62
Activations Density 0.148%