INDEX

Explanations

technical role suffixes (-ifier)

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

-2.47

蹋

-2.42

 scow

-2.42

 gefüllt

-2.33

 مسلمان

-2.31

ㆆ

-2.30

 stalwart

-2.28

 incred

-2.27

 unspeak

-2.25

POSITIVE LOGITS

おすすめの

2.53

手に入

2.28

 traducir

2.13

言うと

2.08

);

2.08

だけど

2.06

 furthermore

2.06

ে

2.00

鶘

1.98

 contemplating

1.97

Activations Density 0.006%