INDEX

Explanations

they

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 térm

-0.63

Was

-0.50

 sådan

-0.50

Was

-0.50

to

-0.49

to

-0.48

각

-0.48

able

-0.47

Taken

-0.47

 foran

-0.46

POSITIVE LOGITS

DockStyle

1.05

<bos>

1.02

Personendaten

0.96

 nakalista

0.84

 shouldn

0.81

awtextra

0.78

SBATCH

0.77

!*\

0.76

UrlResolution

0.76

XmlAccessType

0.76

Activations Density 0.257%