INDEX

Explanations

anything that, should do, anyone wants

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

acz

-0.11

 whatsoever

-0.10

gia

-0.08

 anywhere

-0.08

kus

-0.08

cco

-0.08

 anything

-0.08

 apparently

-0.08

 Shirley

-0.08

 mill

-0.08

POSITIVE LOGITS

 deserves

0.15

 deserve

0.13

 deserving

0.11

 should

0.11

 ought

0.11

should

0.11

 deserved

0.10

 Bowman

0.10

iri

0.10

ua

0.10

Activations Density 0.014%