INDEX

Explanations

multiplexing/twinned words

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

 Kahn

-0.10

mant

-0.09

 Kiss

-0.09

 kiss

-0.09

imit

-0.09

emb

-0.08

 Bates

-0.08

 Higgins

-0.08

 striking

-0.08

 Mason

-0.08

POSITIVE LOGITS

ubb

0.11

 liberalism

0.11

track

0.10

 smarty

0.10

ocalypse

0.10

>window

0.10

 Democr

0.10

mud

0.10

 stubborn

0.10

 Morm

0.10

Activations Density 0.004%