INDEX

Explanations

full action or state

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

or

-1.98

 only

-1.84

 partially

-1.46

at

-1.30

it

-1.28

can

-1.27

 helped

-1.21

 partly

-1.21

 potentially

-1.21

 occasionally

-1.20

POSITIVE LOGITS

and

1.38

 모든

1.30

 любые

1.28

 بكل

1.23

 تمامی

1.22

forall

1.19

どんな

1.18

regardless

1.18

+++

1.17

extremely

1.17

Activations Density 0.060%