INDEX

Explanations

stopping or survival

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

 assisting

-0.86

 helping

-0.85

 preventing

-0.84

 disrupting

-0.76

 Preventing

-0.76

 aiding

-0.75

 interrupting

-0.71

 contributing

-0.71

 guiding

-0.69

 encouraging

-0.69

POSITIVE LOGITS

 Majefty

0.91

ſelf

0.86

 myſelf

0.84

 beſt

0.84

 purpoſe

0.83

 greateſt

0.83

 Diſ

0.82

 reaſon

0.82

 leaſt

0.81

 Houſe

0.79

Activations Density 0.136%