INDEX
Explanations
specific years, particularly 2012 and related dates
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
308
+0.11
0.6%
51
+0.10
0.5%
508
+0.10
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
51
+0.11
0.06
471
+0.10
0.04
151
+0.10
0.04
Negative Logits
Ĺ
-2.73
ĥ
-2.63
ħ
-2.59
·
-2.44
ĺ
-2.34
ĻĤ
-2.31
¶
-2.17
↵
-2.15
↵
-2.15
↵
-2.15
POSITIVE LOGITS
å¹´
2.12
gio
1.76
izumab
1.75
Budget
1.65
nament
1.65
dered
1.56
anniversary
1.56
–
1.56
onwards
1.55
ennial
1.52
Activations Density 0.166%