INDEX

Explanations

Imagine: followed by scenario

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

helpful

1.43

 general

1.31

等方面

1.30

 основном

1.28

主な

1.27

 увагу

1.27

有一定的

1.25

 주로

1.24

 beneficial

1.24

appui

1.24

POSITIVE LOGITS

 crammed

2.46

就这样

2.40

 imagine

2.40

osseum

2.40

 untreated

2.38

明明

2.34

imagine

2.32

 Imagine

2.30

 imagining

2.29

Imagine

2.27

Activations Density 0.462%