INDEX
Explanations
drooping, open, off, empty, shrouded, false
New Auto-Interp
Negative Logits
постепен
0.68
erfolgen
0.64
kształ
0.63
similarity
0.63
stabilit
0.63
والتي
0.63
chaleur
0.62
지속
0.61
त्वरित
0.61
安心して
0.61
POSITIVE LOGITS
intact
1.10
tilted
1.04
partially
1.03
missing
0.99
locked
0.98
empty
0.97
untouched
0.96
turned
0.95
positioned
0.94
stuck
0.94
Activations Density 0.657%