INDEX

Explanations

pronoun followed by verb

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

oft

-0.14

 continuously

-0.13

however

-0.12

 unfortunately

-0.12

 suddenly

-0.12

 slowly

-0.12

aji

-0.12

 myster

-0.12

 however

-0.11

 firstly

-0.11

POSITIVE LOGITS

 scarcely

0.11

 caric

0.10

 neither

0.10

 sober

0.10

Cav

0.10

 glimps

0.09

 grop

0.09

cav

0.09

 cannot

0.09

 Hoff

0.09

Activations Density 0.136%