INDEX

Explanations

a bit of a

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

å¤§å¦Ī

-0.24

æĢ§çĸ¾çĹħ

-0.24

äºĨå¤§éĩıçļĦ

-0.23

ä¹ĭæīĢ

-0.23

ç±»äº§åĵģ

-0.23

ä¸įä»ħèĥ½

-0.23

ä¸įæĸŃå¢ŀåĬł

-0.23

çļĦæĹ¶éĹ´éĩĮ

-0.23

_of

-0.22

lain

-0.22

POSITIVE LOGITS

“

0.28

 luck

0.27

 romantic

0.25

0.24

 overlap

0.24

 compromise

0.23

 balance

0.23

 escape

0.23

 debate

0.22

Activations Density 0.008%