INDEX

Explanations

especially also instance apparently

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

 utilizing

-0.19

 utilized

-0.18

 utilizes

-0.17

 atleast

-0.16

 whilst

-0.16

 mediums

-0.16

Whilst

-0.16

 anyways

-0.15

 advancements

-0.15

POSITIVE LOGITS

 unusually

0.15

 ordinarily

0.15

 chiefly

0.15

 occasional

0.14

 elsewhere

0.13

 partly

0.13

 modest

0.13

 apparently

0.13

 scarcely

0.13

 occasion

0.12

Activations Density 4.529%