INDEX

Explanations

scientific citation patterns

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

مارات

-0.94

in

-0.89

 también

-0.84

Geograf

-0.82

 rekke

-0.81

 already

-0.81

eventType

-0.80

 bereits

-0.79

 which

-0.79

 seem

-0.77

POSITIVE LOGITS

ж

0.93

OGUE

0.89

 krá

0.88

fragt

0.86

zeugt

0.84

ophore

0.81

ガリ

0.81

 coste

0.80

 Vinh

0.79

ovať

0.78

Activations Density 0.015%