INDEX

Explanations

code syntax elements like tags or comments

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

 always

-1.27

かかる

-1.24

 often

-1.20

 herida

-1.18

知ら

-1.14

visiae

-1.12

 águas

-1.08

 erstes

-1.06

otide

-1.04

まるで

-1.04

POSITIVE LOGITS

㉗

1.29

merkmale

1.17

न्ह

1.13

IKAN

1.13

for

1.12

íamos

1.11

zufügen

1.10

回戦

1.09

国道

1.08

 criter

1.07

Activations Density 0.037%