INDEX

Explanations

SELECT * FROM

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

 dipping

1.36

 essences

1.35

 staples

1.30

 illiterate

1.28

 dozens

1.26

 foothills

1.23

 innumerable

1.22

 gaseous

1.22

 equivalents

1.21

য়েছে

1.21

POSITIVE LOGITS

 מה

1.75

 للم

1.72

 لل

1.71

 على

1.65

 رام

1.64

 عن

1.62

 параметра

1.60

 في

1.59

ButtonClick

1.58

 בא

1.55

Activations Density 0.000%