INDEX
Explanations
terms related to synchronization or timing
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
950
+0.14
0.5%
120
+0.13
0.5%
814
+0.12
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
950
+0.14
0.03
120
+0.13
0.02
814
+0.12
0.01
Negative Logits
kosme
-0.68
buta
-0.67
alkoh
-0.67
Bí
-0.65
Bater
-0.64
baf
-0.63
mami
-0.62
interse
-0.62
keramik
-0.62
kac
-0.62
POSITIVE LOGITS
sync
1.24
sync
1.20
Sync
1.11
synchronization
1.11
synchron
1.10
Sync
1.10
SYNC
1.04
syncing
1.04
synchronize
1.02
synchron
1.00
Activations Density 0.112%