INDEX

Explanations

choice is yours

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

enes

-0.10

 Kauf

-0.09

Nel

-0.09

rese

-0.09

 nail

-0.09

Hil

-0.09

 Madd

-0.09

iche

-0.08

lesh

-0.08

 Leipzig

-0.08

POSITIVE LOGITS

 choice

0.30

 Choice

0.24

choice

0.22

 responsibility

0.22

Choice

0.22

 decision

0.21

 depends

0.20

 depend

0.20

-choice

0.18

 depended

0.18

Activations Density 0.097%