INDEX

Explanations

scary spooky

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

ly

-0.94

ener

-0.64

LY

-0.63

eners

-0.61

fully

-0.56

MemoryWarning

-0.54

SourceChecksum

-0.54

LabelTagHelper

-0.52

ništ

-0.52

ually

-0.52

POSITIVE LOGITS

SBATCH

0.58

organ

0.52

izable

0.52

valid

0.52

 Silly

0.52

situ

0.51

 valid

0.50

 <<<<<<<<<<<<<<

0.50

ardin

0.49

OuterClass

0.49

Activations Density 0.007%