INDEX

Explanations

Mixed scraped text

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

=input

-0.27

 ráº¥t

-0.25

ç¿¼

-0.24

ãģ¯ãģļ

-0.24

ç¼ĵæħ¢

-0.24

 resolve

-0.24

éĴĹ

-0.24

æĺ¯éĿŀå¸¸

-0.24

".

-0.24

ed

-0.23

POSITIVE LOGITS

abcdef

0.25

æīĺç¦ı

0.24

gee

0.24

ableObject

0.24

 paramName

0.24

æīĺ

0.24

åĲĮå¹´

0.24

æĲ¦

0.23

FTA

0.23

éĢģåĩº

0.23

Activations Density 0.005%