INDEX

Explanations

I mean to explain or clarify

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pane

-0.27

 luáºŃn

-0.24

çļĦä¸Ģé¡¹

-0.23

urn

-0.21

ounder

-0.21

çļĦä¸Ģå¤§

-0.21

åĲįåı«

-0.21

-guid

-0.21

undra

-0.20

ande

-0.20

POSITIVE LOGITS

 really

0.28

 kidding

0.26

 imagine

0.25

 simply

0.24

 technically

0.24

å¦Ĥæŀľä½ł

0.24

if

0.24

wh

0.23

æ¯ķç«Ł

0.23

Activations Density 0.006%