INDEX
Explanations
names of specific individuals or organizations involved in news events
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
1343
+0.19
0.6%
1499
+0.14
0.4%
304
+0.14
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
981
+0.19
0.09
227
+0.14
0.07
304
+0.14
0.02
Negative Logits
cushi
-1.07
purplish
-1.03
friable
-0.99
scrat
-0.98
glandular
-0.98
disreg
-0.91
nadzie
-0.91
tupperware
-0.91
myce
-0.88
tubercle
-0.87
POSITIVE LOGITS
alkoh
1.55
kompati
1.54
Kategor
1.48
Strukt
1.40
minimalis
1.38
kön
1.37
praktik
1.35
radikal
1.34
keramik
1.34
kosme
1.34
Activations Density 0.311%