INDEX
Explanations
This neuron consistently fires on occurrences of the substring “Euro” (whether referring to the currency or as part of “Euro-X” names).
New Auto-Interp
Negative Logits
!I
-0.08
.tag
-0.07
plant
-0.07
[line
-0.07
logic
-0.07
_type
-0.07
Knowledge
-0.07
риг
-0.07
atan
-0.07
代
-0.07
POSITIVE LOGITS
Euro
0.12
euro
0.11
Euro
0.10
EURO
0.10
euros
0.09
Euros
0.08
Rubio
0.08
€
0.07
electron
0.07
Volvo
0.07
Activations Density 0.005%