INDEX

Explanations

in ... closer

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

 brought

-0.10

IO

-0.10

Cop

-0.09

ascal

-0.09

flat

-0.09

 Starr

-0.09

wyn

-0.09

 fingertips

-0.09

va

-0.09

 blessed

-0.09

POSITIVE LOGITS

 closer

0.18

 close

0.13

 further

0.12

 tighter

0.12

hab

0.10

 clos

0.10

Clo

0.10

 deeper

0.09

 wards

0.09

 towards

0.09

Activations Density 0.003%