INDEX

Explanations

on

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

AndEndTag

-0.77

 continúas

-0.65

chtigkeit

-0.61

(___

-0.61

WriteTagHelper

-0.57

qrstuvwxyz

-0.57

oader

-0.57

chmal

-0.57

ViewFeatures

-0.57

POSITIVE LOGITS

findpost

0.52

set

0.50

 <>",

0.44

 pregnant

0.44

 blood

0.44

 cancer

0.42

jectable

0.42

 asleep

0.42

esthes

0.42

 pills

0.40

Activations Density 0.031%