INDEX
Explanations
incorrect results or truncation
New Auto-Interp
Negative Logits
事業
0.44
사업
0.41
猟
0.40
注重
0.40
husbandry
0.40
ড়াই
0.39
꾼
0.38
วัฒ
0.38
laude
0.37
鹤
0.37
POSITIVE LOGITS
incorrect
0.68
incorrect
0.66
incorrectly
0.65
errone
0.58
Incorrect
0.57
erroneously
0.57
distorted
0.53
ambigu
0.52
falsely
0.50
mismatched
0.50
Activations Density 0.092%