INDEX

Explanations

intensifying adjectives

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

𖣔

-1.27

を

-1.26

 juſ

-1.26

 ASSISTANCE

-1.24

caping

-1.23

鹩

-1.23

uded

-1.23

 granito

-1.20

 vraiment

-1.20

より

-1.18

POSITIVE LOGITS

 important

1.52

 valuable

1.38

 temprano

1.28

 любого

1.22

 rare

1.17

 cuyas

1.16

 valioso

1.16

 cuidadosamente

1.15

 high

1.14

After

1.13

Activations Density 0.033%