INDEX
Explanations
words related to educational collaboration and academic announcements
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1577
+0.17
0.5%
609
+0.14
0.4%
1177
+0.11
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
147
+0.17
0.03
1499
+0.14
0.07
1086
+0.11
0.04
Negative Logits
emphat
-0.74
obstin
-0.67
disgra
-0.67
loup
-0.67
ineffec
-0.64
laft
-0.63
cryst
-0.63
HERBERT
-0.62
capitaine
-0.62
catast
-0.62
POSITIVE LOGITS
setVerticalGroup
0.65
betweenstory
0.64
will
0.56
will
0.56
sẽ
0.55
gradualmente
0.55
TagMode
0.51
StructEnd
0.49
Tazama
0.49
tomorrow
0.48
Activations Density 0.783%