INDEX

Explanations

quality followed by noun

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

but

-1.64

 alguma

-1.40

henge

-1.35

 since

-1.34

 algumas

-1.30

 spesielt

-1.29

But

-1.28

ﻓ

-1.27

-}

-1.25

+{

-1.23

POSITIVE LOGITS

 quality

1.92

 каче

1.70

quality

1.60

 QUALITY

1.55

ve

1.31

↵↵↵↵

1.30

Quality

1.30

 kalite

1.27

 calidad

1.26

↵↵↵

1.24

Activations Density 0.044%