INDEX

Explanations

actions and their results/properties

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

ical

-0.14

ICAL

-0.11

isin

-0.10

ized

-0.10

icans

-0.10

hof

-0.10

culus

-0.10

IZE

-0.10

ality

-0.10

icity

-0.10

POSITIVE LOGITS

ings

0.32

able

0.30

ables

0.22

ers

0.22

ÂŃing

0.19

ability

0.19

ments

0.19

INGS

0.18

ABLE

0.17

ingly

0.16

Activations Density 4.013%