INDEX

Explanations

gather

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

 gather

-1.95

 gathers

-1.84

 gathered

-1.71

 Gather

-1.45

gather

-1.43

Gather

-1.30

Secretary

-1.27

 Secretary

-1.23

 collect

-1.13

collect

-1.13

POSITIVE LOGITS

up

0.60

 back

0.50

0.49

 proof

0.46

 flux

0.46

 routine

0.45

pos

0.44

 flag

0.44

 ultra

0.44

arnes

0.44

Activations Density 0.293%