INDEX

Explanations

buried deep inside

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

 Absolute

-0.09

 regularly

-0.09

 stored

-0.08

 Fernando

-0.08

Bah

-0.08

 Pony

-0.08

 Splash

-0.08

 balance

-0.08

rett

-0.08

 corrid

-0.08

POSITIVE LOGITS

 deeper

0.15

 bury

0.13

 deep

0.13

 treasures

0.12

 deeply

0.12

 treasure

0.11

deep

0.11

 deepest

0.11

 Hatch

0.10

 deepen

0.10

Activations Density 0.021%