INDEX

Explanations

names and subsequent word

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

давно

-1.15

 tyto

-1.15

施行

-1.09

modellen

-1.06

 down

-1.06

 özel

-1.05

</h1>

-1.05

but

-1.05

it

-1.03

 defect

-1.03

POSITIVE LOGITS

 जबकि

1.09

Вот

1.04

 prunes

1.02

CloseOperation

0.96

 României

0.96

 Enough

0.95

フライパン

0.93

 pernic

0.93

 Quince

0.91

 direto

0.91

Activations Density 0.061%