INDEX
Explanations
information related to technical specifications, specifically resolution
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.15
0.9%
1065
+0.12
0.7%
313
+0.12
0.7%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1065
+0.15
0.02
892
+0.12
0.02
851
+0.12
0.02
Negative Logits
<bos>
-2.91
/***
-0.71
InjectMocks
-0.69
promul
-0.68
reinstate
-0.67
amass
-0.65
///**
-0.63
/**
-0.62
reunite
-0.61
lighten
-0.61
POSITIVE LOGITS
resolution
1.28
resolutions
1.24
Resolution
1.22
Resolutions
1.14
resolution
1.12
Resolution
1.12
RESOLUTION
1.08
Resolved
1.02
resolu
1.01
Resolver
0.98
Activations Density 0.078%