INDEX

Explanations

confusion

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Confused

-1.07

 confuse

-1.01

 confuses

-0.99

 confused

-0.99

Confused

-0.95

confused

-0.85

 confound

-0.83

 confounded

-0.79

 confund

-0.75

 bewildered

-0.66

POSITIVE LOGITS

Personendaten

0.68

ActionMode

0.64

تقاوى

0.61

AddTagHelper

0.60

 EconPapers

0.60

ingly

0.55

 sprawie

0.52

 Normdatei

0.51

 عالية

0.50

 typelib

0.49

Activations Density 0.005%

No Known Activations