INDEX
Explanations
SELECT, WHERE, Value, ranges
imperative action commands in code or technical instructions, especially those manipulating windows or performing drawing/movement operations.
This neuron detects mentions of dialog slot names together with their values (e.g. “area = east”, “people = 3”).
New Auto-Interp
Negative Logits
。
0.16
！！！
0.15
\
0.14
!!!!!!!
0.14
,(
0.14
"+"
0.14
"
0.14
AND
0.14
"#
0.13
!!!!!!!!!!!!!!!!
0.13
POSITIVE LOGITS
citado
0.19
quatro
0.19
dépenses
0.18
chefe
0.18
apareció
0.18
സ്വദേശ
0.18
Spanien
0.18
novos
0.18
daños
0.18
romanzo
0.18
Activations Density 0.101%