INDEX

Explanations

delusions or grandiosity

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 psychology

-0.11

 psychologically

-0.11

 competit

-0.09

 inclusive

-0.09

 unhealthy

-0.09

 unknow

-0.09

 impartial

-0.09

 courageous

-0.09

zc

-0.09

 Psychology

-0.09

POSITIVE LOGITS

 charm

0.13

 Charm

0.13

 rage

0.12

 Grand

0.12

 charming

0.11

 grand

0.11

 brag

0.10

Grand

0.10

grand

0.10

 Lazar

0.10

Activations Density 0.005%

No Known Activations