INDEX

Explanations

showing

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

éĻĢ

-0.28

çģĮ

-0.26

 bottles

-0.25

ivas

-0.25

imers

-0.25

çıĢ

-0.25

_ROT

-0.24

ham

-0.24

æĹĭ

-0.24

Ð½Ð°ÑĤ

-0.24

POSITIVE LOGITS

 showing

0.57

 shows

0.53

 show

0.52

show

0.48

æĺ¾ç¤º

0.48

 Shows

0.46

shows

0.45

	show

0.45

 Showing

0.43

æĺ¾ç¤ºåĩº

0.43

Activations Density 0.925%