INDEX
Explanations
sentences ending in periods that describe actions or situations
sentences that indicate the conclusion of a statement
New Auto-Interp
Negative Logits
intermediate
-0.65
assignments
-0.63
predec
-0.62
wherever
-0.60
joints
-0.60
parity
-0.60
biological
-0.58
homework
-0.58
multiplication
-0.58
premium
-0.57
POSITIVE LOGITS
REUTERS
1.02
SPONSORED
0.98
Hide
0.94
Photograph
0.87
Critics
0.85
Photo
0.84
Their
0.82
Caption
0.82
PHOTO
0.80
Downloadha
0.80
Activations Density 0.379%