INDEX

Explanations

bragging, boasting, or ranting

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

avis

-0.09

unya

-0.08

 mole

-0.08

 uphill

-0.08

vae

-0.08

atings

-0.08

 downhill

-0.08

PA

-0.07

PA

-0.07

Hor

-0.07

POSITIVE LOGITS

 swagger

0.14

 brag

0.13

-worthy

0.12

 boasting

0.11

 pled

0.11

asher

0.10

-about

0.10

swagger

0.10

evin

0.10

 Blind

0.10

Activations Density 0.003%