INDEX
Explanations
spatial concentration or clustering
New Auto-Interp
Negative Logits
:+:
0.42
напрямую
0.40
সম্পূর্ণরূপে
0.40
场合
0.40
직접
0.39
subcategory
0.38
可以直接
0.38
ault
0.37
Tin
0.37
媲
0.37
POSITIVE LOGITS
集中
1.50
concentrated
1.48
clustered
1.41
clustering
1.36
clusters
1.33
koncent
1.25
clustering
1.21
Clustering
1.21
concent
1.20
cluster
1.16
Activations Density 0.040%