INDEX

Explanations

legacy and past systems

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

emic

-0.09

hoot

-0.08

otal

-0.08

lic

-0.08

 tongue

-0.08

avit

-0.08

ICO

-0.08

ous

-0.08

line

-0.08

eer

-0.08

POSITIVE LOGITS

/Grid

0.12

(trans

0.11

-padding

0.11

cai

0.11

-shell

0.10

Nad

0.10

 cara

0.10

 matrimon

0.10

fully

0.10

ehr

0.10

Activations Density 0.011%