INDEX

Explanations

still followed by verb

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 RESTAUR

-1.27

蟾

-1.23

έχ

-1.19

ższe

-1.13



-1.12

icherheit

-1.11

 RECIPES

-1.10

 recupero

-1.10

 seguro

-1.09

 CENTRAL

-1.08

POSITIVE LOGITS

by

1.33

 bazı

1.27

 Asegúrese

1.20

 Apenas

1.20

 Puedes

1.20

 BEST

1.16

tyl

1.13

pædia

1.12

 STILL

1.10

 ако

1.10

Activations Density 0.063%