INDEX

Explanations

treatment and restoration

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

 still

-1.23

if

-1.02

 ontwikkeling

-0.96

 Vorgang

-0.92

still

-0.89

setzungen

-0.87

indest

-0.86

 gemakkelijk

-0.86

💆

-0.84

 elkaar

-0.83

POSITIVE LOGITS

 Finally

1.10

 finally

1.09

 improved

1.02

SOLUTION

1.01

 dopiero

1.00

 اصلاح

0.99

去除

0.98

OPSIS

0.96

後來

0.95

찾

0.94

Activations Density 0.077%