INDEX
Explanations
The main thing this neuron does is find names related to various political figures or individuals
names and references to individuals, particularly those with the surname "McGuinness."
New Auto-Interp
Negative Logits
è£ı
-0.75
ãĤ¼
-0.74
Gemini
-0.74
ãĥ³ãĤ¸
-0.74
çīĪ
-0.71
è£ıç
-0.69
Trojan
-0.69
CBI
-0.69
ãĥ¼ãĥ³
-0.68
Saiyan
-0.67
POSITIVE LOGITS
inness
1.27
igans
1.02
McGu
0.96
igan
0.84
olicy
0.84
pload
0.83
issance
0.82
ire
0.81
essing
0.80
idge
0.79
Activations Density 0.018%