INDEX

Explanations

Q, A, Task beginnings

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

 raso

-0.77

 valla

-0.76

moud

-0.73

 karo

-0.73

общение

-0.72

Offer

-0.72

 justifyContent

-0.71

 valo

-0.70

 қара

-0.70

需要

-0.68

POSITIVE LOGITS

 Wyoming

1.01

 Juve

1.00

 Cheyenne

0.92

Wyoming

0.91

 Yellowstone

0.84

PropertyChanging

0.82

 semilla

0.82

 atentos

0.81

 Péter

0.81

 Fracture

0.81

Activations Density 0.002%