INDEX

Explanations

YouTube search queries

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 위하여

2.42

 平成

2.42

 不能

2.36

 但是

2.32

秶

2.30

也

2.25

乀

2.24

 변경

2.23

又

2.19

 것을

2.18

POSITIVE LOGITS

 badass

2.31

—

1.93

 behem

1.91

 gigantes

1.82

 reimag

1.81

<em>

1.79

 incredibly

1.78

 pesky

1.73

 standout

1.72

 dudes

1.71

Activations Density 0.409%