INDEX

Explanations

conditional logic or choices

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

ában

-0.91

 historien

-0.84

üht

-0.83

 caract

-0.82

 erit

-0.82

 électron

-0.81

 halle

-0.80

艏

-0.79

 annealing

-0.79

burgs

-0.79

POSITIVE LOGITS

 depending

1.40

 dependiendo

1.17

 accordance

1.06

 megfelelő

0.99

depending

0.95

due

0.94

 зависит

0.92

 prefer

0.90

 devido

0.88

 根据

0.88

Activations Density 0.069%