INDEX

Explanations

carry flags

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

 initiating

-0.77

仲

-0.73

альные

-0.73

onents

-0.73

Critique

-0.71

 during

-0.70

ា

-0.69

メニューは

-0.68

atex

-0.68

 since

-0.68

POSITIVE LOGITS

 flags

1.25

 condition

1.20

Flags

1.19

 Carry

1.17

 Zero

1.16

 flag

1.16

Zero

1.16

Carry

1.15

 carry

1.15

Condition

1.13

Activations Density 0.022%