INDEX
Explanations
This neuron fires on terms referring to aerial or bird’s‐eye viewpoints (e.g., “aerial,” “drone,” “bird‐eye,” “view,” “survey”).
New Auto-Interp
Negative Logits
valu
-0.06
.testng
-0.06
Mon
-0.06
\Factory
-0.06
Pregn
-0.06
_for
-0.06
Provid
-0.06
asl
-0.06
allest
-0.06
CLOSED
-0.06
POSITIVE LOGITS
aerial
0.09
erial
0.08
الإ
0.07
inclined
0.07
亭
0.07
多い
0.07
436
0.07
DEBUG
0.06
NECT
0.06
vrch
0.06
Activations Density 0.005%