INDEX
Explanations
Conversational language
The neuron activates on words that refer to personal acquaintances or informal sources of help (e.g. “anyone,” “buddy,” “run a newsagency”).
New Auto-Interp
Negative Logits
############################################################################
-0.07
Grab
-0.06
_suite
-0.06
Historical
-0.06
waiting
-0.06
Emerging
-0.06
registration
-0.06
porcelain
-0.06
friendly
-0.06
ondon
-0.06
POSITIVE LOGITS
pany
0.07
.XPATH
0.06
GLUT
0.06
SAN
0.06
JNIEnv
0.06
lãi
0.06
بات
0.06
newline
0.06
vale
0.06
sale
0.06
Activations Density 0.029%