INDEX

Explanations

closing code blocks and strings

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

-1.82

of

-1.57

-1.54

〜。

-1.40

at

-1.34

by

-1.27

〜！

-1.25

轍

-1.16

 before

-1.12

 using

-1.11

POSITIVE LOGITS

に入れる

1.34

,</

1.28

atization

1.28

upported

1.27

をご覧

1.27

 fundas

1.27

ParameterLimits

1.26

ふん

1.26

感觉到

1.26

atics

1.26

Activations Density 0.006%