INDEX

Explanations

Supreme Court

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Simpsons

-0.12

 Hermione

-0.11

 Kore

-0.11

 refrigerator

-0.10

 dove

-0.10

hsv

-0.10

 pasta

-0.10

 Simpson

-0.10

Biz

-0.10

 Andrea

-0.10

POSITIVE LOGITS

 Jack

0.50

Jack

0.43

 jack

0.43

jack

0.39

 JACK

0.38

 Jackie

0.31

 jackpot

0.30

jak

0.18

 Jackson

0.18

Jak

0.17

Activations Density 0.069%

No Known Activations