INDEX

Explanations

direction

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

SEGUIR

-0.60

 Offisielt

-0.59

 nôtre

-0.57

 chré

-0.57

czeń

-0.57

 voyons

-0.56

 laikā

-0.56

 conseguenze

-0.55

 jaunes

-0.54

 poupée

-0.52

POSITIVE LOGITS

 were

0.56

were

0.56

EndGlobalSection

0.53

where

0.53

 torr

0.53

Personensuche

0.52

 where

0.52

lántico

0.52

providedIn

0.52

OCCURRED

0.51

Activations Density 0.101%