INDEX

Explanations

extent and size

abstract noun endings

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ционной

0.53

ционного

0.42

ционный

0.42

ალურ

0.42

шлай

0.42

FinishLaunching

0.41

тического

0.41

čního

0.41

ATORY

0.41

ATIONS

0.40

POSITIVE LOGITS

ness

1.40

ность

1.18

NESS

1.16

ність

1.13

ność

1.12

eness

1.05

ość

1.05

ότητα

1.05

さを

1.02

kość

1.02

Activations Density 0.018%