INDEX

Explanations

Pronouns

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

casio

-0.44

Personendaten

-0.43

Démographie

-0.37

RefNanny

-0.36

FactoryBean

-0.35

 tắc

-0.34

 cáncer

-0.33

bau

-0.33

 pantaloni

-0.33

bett

-0.32

POSITIVE LOGITS

 being

1.50

being

1.31

 having

1.30

having

1.22

Being

1.12

 BEING

1.06

 étant

1.03

 Being

1.01

Having

1.00

 HAVING

1.00

Activations Density 0.001%

No Known Activations