INDEX
Explanations
legal terms and conditions related to website usage
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.29
1.2%
1403
+0.09
0.4%
453
+0.08
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1403
+0.29
0.05
1419
+0.09
0.07
248
+0.08
0.06
Negative Logits
<bos>
-1.82
ⓧ
-1.04
/**
-0.84
-0.83
<?
-0.70
Transcripción
-0.68
<?
-0.68
overthrown
-0.67
disbur
-0.65
shivered
-0.64
POSITIVE LOGITS
pleins
1.18
(%)
1.00
éto
0.96
↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
0.94
↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
0.93
mavi
0.92
vasi
0.92
usak
0.92
catég
0.92
-)
0.91
Activations Density 1.055%