INDEX

Explanations

RK or rg followed by specific tokens

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

Cong

-0.10

ancial

-0.10

 Cong

-0.10

 Conj

-0.10

ren

-0.09

rier

-0.09

erator

-0.09

 unsub

-0.09

 widow

-0.09

ickle

-0.08

POSITIVE LOGITS

 Ingram

0.10

 racing

0.09

angelo

0.09

ëıĦ

0.09

ieten

0.09

Pru

0.09

olta

0.09

rac

0.08

fal

0.08

oton

0.08

Activations Density 0.086%