INDEX

Explanations

spatial locations and parts

the top, the back

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

篒

0.41

ensos

0.40

 وابسته

0.38

ႃႇ

0.37

蜚

0.37

 бес

0.35

 जबरदस्त

0.34

 धमाकेदार

0.34

 аўтаматы

0.34

 خاصيه

0.34

POSITIVE LOGITS

 bottom

1.66

 upper

1.51

 underside

1.45

bottom

1.42

底部

1.42

top

1.39

 lower

1.34

 middle

1.34

upper

1.30

 sides

1.30

Activations Density 0.742%