INDEX

Explanations

gold followed by names or descriptors

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

er

-0.98

et

-0.93

ians

-0.93

 operativo

-0.87

ir

-0.85

ed

-0.80

LL

-0.80

all

-0.78

iri

-0.78

 Term

-0.76

POSITIVE LOGITS

 plated

1.35

plated

1.28

mine

1.16

 medal

1.15

richtig

1.14

ilo

1.14

 Plated

1.13

panning

1.09

gold

1.05

 foil

1.02

Activations Density 0.016%