INDEX
Explanations
references to age, specifically focusing on older individuals
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
362
+0.15
0.8%
156
+0.14
0.7%
128
+0.12
0.6%
Correlated Neurons
Index
P. Corr.
Cos Sim.
362
+0.15
0.03
128
+0.14
0.02
1
+0.12
0.02
Negative Logits
lest
-1.73
ights
-1.73
async
-1.57
alysis
-1.53
ycl
-1.49
tte
-1.46
esan
-1.46
oof
-1.45
asek
-1.45
ount
-1.41
POSITIVE LOGITS
than
2.60
generations
2.18
Than
2.13
Than
2.04
versions
1.93
generation
1.88
than
1.76
adulthood
1.67
version
1.60
(âī¥
1.59
Activations Density 0.056%