INDEX
Explanations
can/could followed by action
New Auto-Interp
Negative Logits
evita
0.42
দিয়
0.41
evolucion
0.41
иногда
0.40
பர
0.40
อ่ะ
0.39
नंतर
0.38
gü
0.38
eventualmente
0.38
accep
0.38
POSITIVE LOGITS
scarcely
0.67
rightly
0.64
rightfully
0.64
boast
0.61
attest
0.57
rightful
0.54
hardly
0.53
justly
0.50
boasted
0.47
expect
0.47
Activations Density 0.012%