INDEX
Explanations
situations or elements related to geopolitical events
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
645
+0.10
0.3%
1044
+0.08
0.2%
1438
+0.08
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
645
+0.10
0.04
172
+0.08
0.03
789
+0.08
0.02
Negative Logits
reluct
-1.83
disagre
-1.78
increa
-1.76
shenan
-1.73
encomp
-1.69
impra
-1.68
depic
-1.68
indestru
-1.67
unspeak
-1.64
inconce
-1.63
POSITIVE LOGITS
useCallback
0.69
also
0.64
also
0.64
weil
0.64
também
0.62
importantly
0.62
también
0.61
ook
0.60
också
0.60
ALSO
0.60
Activations Density 0.183%