INDEX
Explanations
actions or tasks related to taking care of someone's hygiene and daily needs
It detects tokens that introduce lists or descriptions of included activities, services, or course content.
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.13
0.4%
752
+0.10
0.3%
369
+0.10
0.3%
Correlated Neurons
Index
P. Corr.
Cos Sim.
752
+0.13
0.05
16
+0.10
0.08
369
+0.10
0.05
Negative Logits
unlaw
-1.09
inappro
-1.06
impractica
-1.06
quitted
-1.01
unve
-0.99
uninten
-0.96
reluct
-0.95
wherea
-0.93
disreg
-0.93
unwarran
-0.91
POSITIVE LOGITS
various
0.92
everything
0.79
både
0.78
both
0.77
numerous
0.73
various
0.71
both
0.68
Various
0.67
sowohl
0.66
Various
0.63
Activations Density 0.792%