INDEX
Explanations
references to injuries and medical issues in sports-related contexts
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
964
+0.09
0.3%
495
+0.08
0.2%
218
+0.07
0.2%
Correlated Neurons
Index
P. Corr.
Cos Sim.
619
+0.09
0.02
495
+0.08
0.03
1449
+0.07
0.01
Negative Logits
frambo
-0.61
ananas
-0.60
SBATCH
-0.56
marte
-0.56
aquare
-0.55
ükemmel
-0.54
herbes
-0.53
grotte
-0.52
stiller
-0.52
svin
-0.52
POSITIVE LOGITS
gaily
0.71
vainly
0.70
previously
0.67
thrived
0.67
originally
0.65
formerly
0.63
quitted
0.63
ineffec
0.62
vanished
0.61
contribut
0.59
Activations Density 0.590%