INDEX

Explanations

research articles and technical descriptions

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

Should

-1.36

（

-1.32

 quindi

-1.28

他的

-1.27

<bos>

-1.25

Without

-1.25

在

-1.23

Here

-1.22

Through

-1.17

leding

-1.15

POSITIVE LOGITS



1.48

are

1.45

dzić

1.38

1.36

|]

1.28

/**

1.28

serve

1.27

FileType

1.25

垸

1.23

 devise

1.23

Activations Density 0.006%