INDEX

Explanations

history

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

è¹¿

-0.28

//*

-0.28

çªľ

-0.27

emat

-0.26

erva

-0.25

åĪĽ

-0.25

char

-0.25

 ener

-0.24

bolt

-0.24

.createUser

-0.24

POSITIVE LOGITS

éĩĩéĽĨ

0.40

æĶ¶éĽĨ

0.39

 acquisition

0.37

æĲľéĽĨ

0.36

èİ·åıĸ

0.34

 gather

0.33

 collection

0.33

åĮħæĭ¬

0.32

 Gather

0.32

 gathering

0.31

Activations Density 0.786%