INDEX

Explanations

reproducible results and time

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

 filenames

1.60

 frameN

1.58

 μόνο

1.55

ᓕ

1.54

 けれど

1.52

 文件

1.52

 කරයි

1.51

 하지만

1.51

 以及

1.49

 функций

1.45

POSITIVE LOGITS

nav

1.10

ruiting

0.98

sc

0.97

Add

0.96

adir

0.95

icorn

0.93

ne

0.93

invest

0.91

ug

0.91

used

0.90

Activations Density 0.022%