INDEX

Explanations

evolution, delay, highly, creates

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

 twist

-0.10

 clutch

-0.10

este

-0.10

vos

-0.10

tou

-0.10

 barely

-0.10

 Bosch

-0.10

ç±

-0.10

 feather

-0.10

 syndrome

-0.09

POSITIVE LOGITS

up

0.10

 Intercept

0.09

PIT

0.09

 down

0.08

 Hind

0.08

Heg

0.08

pit

0.08

 mamma

0.08

 Hond

0.08

 near

0.07

Activations Density 0.001%