INDEX
Explanations
mentioned family relations, particularly grandparents and grandfathers
references to grandparents and their relationships in the text
New Auto-Interp
Negative Logits
ograp
-0.71
ylon
-0.69
tics
-0.69
acus
-0.68
ijk
-0.67
inem
-0.66
airo
-0.66
ographed
-0.65
coord
-0.65
mberg
-0.65
POSITIVE LOGITS
father
0.92
grandmother
0.89
parents
0.87
grandma
0.87
grandparents
0.84
grandfather
0.82
sson
0.82
Takeru
0.79
ENTS
0.78
sie
0.77
Activations Density 0.056%