INDEX
Explanations
conjunctions and transition words
New Auto-Interp
Negative Logits
,
0.74
:
0.64
commonplace
0.57
ים
0.56
rigorously
0.56
perme
0.55
competencies
0.55
ю
0.55
relev
0.54
օ
0.54
POSITIVE LOGITS
Hence
1.10
But
1.09
However
1.02
In
1.00
With
0.99
Therefore
0.99
When
0.98
Because
0.98
And
0.98
While
0.98
Activations Density 0.078%