INDEX
Explanations
This neuron detects mentions of bridges (including names or the word “bridge”/“Brücke”).
New Auto-Interp
Negative Logits
incr
-0.06
TeV
-0.06
hott
-0.06
soul
-0.06
Dump
-0.06
Src
-0.06
Ramos
-0.06
andbox
-0.06
郭
-0.06
sail
-0.06
POSITIVE LOGITS
bridge
0.10
iesz
0.08
"?"
0.07
Bridge
0.07
př
0.07
phép
0.07
여
0.06
(#)
0.06
uating
0.06
.......
0.06
Activations Density 0.010%