INDEX

Explanations

more than double, nearly doubled

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

readFileSync

-1.42

/${

-1.41

”、“

-1.27

)=>{

-1.17

 mengumpulkan

-1.17

.}(

-1.16

 coraz

-1.14

iné

-1.12

 baweł

-1.09

çois

-1.07

POSITIVE LOGITS

一丝

1.23

 first

1.20

特点

1.19

防护

1.19

 last

1.17

 apartament

1.13

 campi

1.13

 Chriftian

1.12

 prega

1.10

Activations Density 0.045%