INDEX

Explanations

something being shown/presented

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

 shown

-0.86

 given

-0.73

 Shown

-0.69

Efq

-0.65

Eſ

-0.62

 whoſe

-0.59

given

-0.58

 occaf

-0.57

 fhew

-0.57

 Reſ

-0.57

POSITIVE LOGITS

yntaxException

0.80

0.55

ed

0.52

 implying

0.52

usercontent

0.51

DATABASES

0.51

arXiv

0.50

ToAction

0.50

 betweenstory

0.50

Tikang

0.49

Activations Density 0.319%