INDEX
Explanations
words related to religion and family, such as religious attendance, religious commitment, importance of religion in daily life, religious affiliation, and self-defense against attack
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
394
+0.27
0.9%
1842
+0.24
0.8%
50
+0.18
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1842
+0.27
0.11
394
+0.24
0.09
284
+0.18
0.09
Negative Logits
<bos>
-1.33
scopri
-0.69
rispond
-0.69
specialmente
-0.67
purtroppo
-0.66
scel
-0.66
potete
-0.63
sappi
-0.63
nemmeno
-0.61
Примеча
-0.60
POSITIVE LOGITS
ErrorCode
0.71
labd
0.70
TokenType
0.69
getUserId
0.68
Inhabitants
0.67
UserData
0.67
AssertionError
0.66
resultList
0.66
siena
0.66
BigNumber
0.65
Activations Density 1.344%