INDEX

Explanations

bite the bullet and move on

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

adan

-0.11

 Fischer

-0.11

ma

-0.10

 instrumental

-0.10

 Anch

-0.10

apan

-0.10

LOY

-0.10

Pes

-0.10

 unde

-0.10

Pax

-0.10

POSITIVE LOGITS

 anymore

0.10

 lets

0.10

lah

0.10

 Torres

0.10

 Beard

0.08

fam

0.08

ness

0.08

 shave

0.08

hed

0.08

 Meredith

0.08

Activations Density 0.001%

No Known Activations