INDEX
Explanations
social/political issues
This neuron fires on mentions of organizations’ funding or support decisions tied to discrimination policies (e.g. withdrawing or withholding money because a group discriminates).
New Auto-Interp
Negative Logits
Result
-0.06
path
-0.06
year
-0.06
Tail
-0.06
真
-0.06
徐
-0.06
Tail
-0.06
써
-0.06
,与
-0.06
Tweets
-0.06
POSITIVE LOGITS
Synopsis
0.06
charAt
0.06
Trotsky
0.06
.setSize
0.06
prostituer
0.06
-auth
0.06
initial
0.06
+'.
0.06
entfer
0.06
erset
0.06
Activations Density 0.040%