INDEX

Explanations

mathematical symbols and numbers

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

to

-1.87

诸多

-1.55

by

-1.48

</b>

-1.45

 However

-1.37

遊ぶ

-1.34

یکر

-1.34

そういった

-1.33

这类

-1.32

皚

-1.28

POSITIVE LOGITS

",

1.68

重要

1.32

肥

1.31

 zawsze

1.30

 anklicken

1.29

 which

1.26

şu

1.24

明白

1.23

',

1.23

آیا

1.23

Activations Density 0.002%