INDEX
Explanations
legal case references and terminology
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
410
+0.22
1.3%
369
+0.17
1.0%
332
+0.14
0.8%
Correlated Neurons
Index
P. Corr.
Cos Sim.
410
+0.22
0.09
369
+0.17
0.04
289
+0.14
0.03
Negative Logits
)',
-1.97
{}-1.89
exports
-1.85
)",
-1.84
oretic
-1.66
ibraries
-1.63
malloc
-1.63
PCs
-1.59
ras
-1.58
etc
-1.58
POSITIVE LOGITS
´
3.13
·
3.13
º
3.12
Į
3.05
¾
3.04
°
3.03
³
2.96
¸
2.95
²
2.93
ľ
2.85
Activations Density 0.352%