INDEX
Explanations
phrases associated with affordability or reasonable pricing
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.08
0.3%
1896
+0.07
0.3%
1506
+0.06
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
123
+0.08
0.03
1506
+0.07
0.03
1640
+0.06
0.03
Negative Logits
<bos>
-1.55
<?
-0.80
.
-0.77
-0.77
//
-0.73
can
-0.73
,
-0.73
addComponent
-0.72
let
-0.71
<eos>
-0.70
POSITIVE LOGITS
affor
2.23
maneu
2.21
impra
2.10
tolerably
2.09
increa
2.04
excru
1.95
stockholm
1.92
accla
1.92
!...
1.91
emphat
1.87
Activations Density 0.108%