INDEX

Explanations

forming a hypothesis

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

первых

-0.55

ynh

-0.55

__((

-0.54

 gynhyrchwyd

-0.54

 ddelweddau

-0.54

Diweddarwch

-0.54

ver

-0.53

 Dyna

-0.52

ViewFeatures

-0.52

 HttpNotFound

-0.50

POSITIVE LOGITS

DoubleQuotes

0.67

 Pind

0.59

 Cæsar

0.57

 Rady

0.57

 Shakspeare

0.57

 Sarm

0.54

 Aphrodite

0.53

 mág

0.52

OSP

0.52

 Diri

0.51

Activations Density 0.261%