INDEX

Explanations

keep our

lists, bullet points, or definitions

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 darken

0.48

 anschließend

0.46

 acceleration

0.45

红色

0.45

ILL

0.44

 suddenly

0.44

acceleration

0.44

리지

0.44

 accelerate

0.43

鳃

0.43

POSITIVE LOGITS

 يحتوي

0.41

 那麼

0.41

 Clearly

0.40

 Katz

0.40

 Zebra

0.40

 unicode

0.39

 True

0.39

 Commitment

0.39

 すご

0.39

kait

0.39

Activations Density 0.002%