INDEX

Explanations

code with types and parentheses

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

-1.41

 first

-1.36

or

-1.35

を選択します

-1.23

and

-1.19

 obok

-1.19

‘

-1.18

the

-1.16

-1.08

づけ

-1.07

POSITIVE LOGITS

もあって

1.47

 degust

1.25

 substan

1.19

はこの

1.07

<bos>

1.07

 conduce

1.05

 hura

1.05

こういった

1.05

 такого

1.04

 그의

1.03

Activations Density 0.003%