INDEX

Explanations

starting from the beginning or end

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etch

-0.11

ITES

-0.11

 throughout

-0.11

 onto

-0.11

 entire

-0.10

omi

-0.10

illin

-0.10

 into

-0.10

Nx

-0.09

into

-0.09

POSITIVE LOGITS

 starting

0.14

 start

0.14

 beginning

0.13

 Starting

0.13

Starting

0.13

starting

0.13

 Beginning

0.13

 bottom

0.13

 begin

0.13

 ìĭľìŀĳ

0.13

Activations Density 0.112%