INDEX

Explanations

as a descriptor

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

 existing

-0.84

drawiam

-0.82

 maillot

-0.79

 Secretaries

-0.79

Attacking

-0.79

 ambiental

-0.78

 tournant

-0.78

Extraction

-0.77

Inves

-0.76

Twe

-0.76

POSITIVE LOGITS

 from

0.91

名

0.87

也

0.84

 Join

0.82

 специаль

0.82

bootstrap

0.82

HANDLE

0.82

in

0.80

Join

0.80

rowIndex

0.79

Activations Density 0.060%