INDEX

Explanations

subset portion

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ainties

-0.22

ytt

-0.20

toPromise

-0.19

irit

-0.18

[]){↵

-0.18

ucci

-0.18

iesen

-0.18

åħ¨å¤©

-0.18

 Farrell

-0.17

/Register

-0.17

POSITIVE LOGITS

 subset

0.47

éĥ¨åĪĨ

0.44

 portion

0.41

subset

0.40

æĪª

0.40

ä¸Ģéĥ¨åĪĨ

0.39

portion

0.39

 Subset

0.38

Subset

0.38

 partial

0.37

Activations Density 0.819%

No Known Activations