INDEX

Explanations

making choices wisely/carefully

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

 Neurol

-0.95

preferred

-0.93

 Το

-0.85

deur

-0.84

pretty

-0.81

Preferred

-0.81

nova

-0.81

igned

-0.81

llow

-0.80

中华

-0.80

POSITIVE LOGITS

 wisely

4.84

 wise

3.34

 Wise

3.25

Wise

2.78

wise

2.66

 carefully

2.58

 unwise

2.44

 wisest

2.42

 wiser

2.30

 WISE

2.17

Activations Density 0.041%