INDEX

Explanations

abbreviations followed by context

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

are

-1.81

-1.54

</h3>

-1.45

It

-1.44

The

-1.34

mathrm

-1.31

安徽

-1.30

It

-1.28

陕西

-1.27



-1.27

POSITIVE LOGITS

 食堂

1.52

頹

1.41

 émotion

1.40

に行きました

1.38

いは

1.31

 oscu

1.31

 सत्यापित

1.31

 nhưng

1.30

 gefertigt

1.30

Reduces

1.29

Activations Density 0.000%