INDEX

Explanations

email notifications and files

The neuron is primarily picking up on double‐quoted string literals (e.g. encoded MIME types, base64 flags, file names, etc.)

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

of

-2.45

is

-1.82

to

-1.47

-1.36

 نظر

-1.33

as

-1.29

 people

-1.28

保健

-1.18

 things

-1.14

we

-1.13

POSITIVE LOGITS

 sonriente

1.57

 gracioso

1.37

官網

1.35

 nắp

1.34

拍了拍

1.32

 your

1.32

λους

1.29

ᅲ

1.27

 sucedido

1.27

Уважа

1.27

Activations Density 0.247%