INDEX

Explanations

positive recognition

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

ADS

-0.08

relay

-0.08

cona

-0.08

oline

-0.07

 rigid

-0.07

lopen

-0.07

 prohibition

-0.07

 attempting

-0.07

 controlled

-0.07

POSITIVE LOGITS

 flattering

0.34

 compliment

0.17

 attractiveness

0.17

ego

0.16

 popularity

0.16

 narciss

0.16

 vanity

0.15

 Recognition

0.15

 attraction

0.15

 recognition

0.15

Activations Density 0.066%