INDEX
Explanations
verbs and technical terms
Tokens that start a new sentence or section — especially capitalized/leading words beginning headings or sentence-initial phrases.
New Auto-Interp
Negative Logits
canceled
0.41
ましたが
0.40
null
0.39
friend
0.39
Aufbau
0.38
හේ
0.37
Saclay
0.37
carbonyl
0.37
assigned
0.37
valamint
0.37
POSITIVE LOGITS
适合
0.45
బె
0.44
ాడు
0.43
만
0.43
덕
0.43
aje
0.43
เหมาะ
0.43
moil
0.42
чески
0.41
дя
0.41
Activations Density 0.000%