INDEX

Explanations

study details and intros

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

 then

-2.09

 know

-1.62

It

-1.41

 suficiente

-1.37

how

-1.34

 gehör

-1.30

 consider

-1.30

it

-1.30

 finding

-1.27

 when

-1.26

POSITIVE LOGITS

 生日

1.45

alı

1.27

훙

1.24

}_{

1.23

taminophen

1.23

\\

1.23

americanos

1.23

↵↵↵↵↵↵↵↵↵↵↵

1.23

 внутренних

1.22

.,

1.21

Activations Density 0.200%