INDEX

Explanations

places followed by 'in'

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

and

-1.83

下記

-1.49

 milyen

-1.48

 hogyan

-1.42

-1.39

 rodillas

-1.37

/***

-1.33

ようです

-1.33

づけ

-1.30

POSITIVE LOGITS

in

3.23

of

2.92

<bos>

1.87

on

1.77

それでも

1.72

參考文獻

1.61

 that

1.58

だからこそ

1.55

個人的には

1.55

 цвят

1.47

Activations Density 0.066%