INDEX

Explanations

even how

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

 hingegen

-0.09

 allerdings

-0.09

 toutefois

-0.08

 majd

-0.08

 inoltre

-0.08

åĬłä¹ĭ

-0.08

ortun

-0.08

adou

-0.07

ambar

-0.07

 Ð¾Ð±Ð°ÑĩÐµ

-0.07

POSITIVE LOGITS

/OR

0.09

 ones

0.09

the

0.09

its

0.08

 those

0.08

 their

0.08

äººæĿ¥è¯´

0.07

 Ø§ÛĮÙĨÚ©Ùĩ

0.07

çļĦäºĨ

0.07

,to

0.07

Activations Density 0.232%