INDEX
Explanations
This neuron is primarily triggered by the common preposition “in.”
New Auto-Interp
Negative Logits
auto
-0.07
Fd
-0.06
푸
-0.06
style
-0.06
sound
-0.06
hexadecimal
-0.06
usuario
-0.06
Derek
-0.06
title
-0.06
說
-0.06
POSITIVE LOGITS
ضو
0.07
ممن
0.07
полож
0.06
кроме
0.06
можлив
0.06
ήν
0.06
%"
0.06
cess
0.06
.setMaximum
0.06
_in
0.06
Activations Density 0.007%