INDEX

Explanations

a great deal of

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

 Bowman

-0.08

 unforgettable

-0.08

ville

-0.08

 tough

-0.07

gage

-0.07

 unparalleled

-0.07

inton

-0.07

 Kear

-0.07

 decent

-0.07

 paradise

-0.07

POSITIVE LOGITS

-grand

0.21

 deal

0.20

deal

0.18

 DEAL

0.15

 Deal

0.14

sword

0.14

 dane

0.13

Deal

0.13

 pains

0.13

 dealt

0.13

Activations Density 0.048%