INDEX
Explanations
phrases related to structural collapses
references to structural failures or collapses
New Auto-Interp
Negative Logits
alty
-0.77
ership
-0.75
lance
-0.72
friend
-0.70
ificantly
-0.70
cipled
-0.70
friends
-0.68
Way
-0.67
Demand
-0.67
chie
-0.66
POSITIVE LOGITS
exting
0.92
tremend
0.87
collapsed
0.86
shells
0.83
lungs
0.82
rubble
0.80
subur
0.80
collapsing
0.76
exha
0.75
withd
0.73
Activations Density 0.027%