INDEX

Explanations

chained, tied, restrained

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

 confisc

-0.11

 sickness

-0.10

/non

-0.10

 slur

-0.10

 holland

-0.09

insky

-0.09

sou

-0.09

 mush

-0.09

 Collapse

-0.09

 Injury

-0.09

POSITIVE LOGITS

 chained

0.19

 tied

0.16

 awaiting

0.15

await

0.13

Await

0.12

 restrained

0.12

 ready

0.12

 struggling

0.12

 hanging

0.12

 undergoing

0.12

Activations Density 0.008%