INDEX

Explanations

door, lock, and entry

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

 exits

-0.14

 exit

-0.13

 exited

-0.13

 exiting

-0.13

 aisle

-0.12

 Exit

-0.12

 EXIT

-0.11

EXIT

-0.11

 Ceiling

-0.11

(exit

-0.11

POSITIVE LOGITS

 porch

0.17

 doorstep

0.13

 Knock

0.13

 knock

0.12

 knocking

0.12

 mailbox

0.11

 sidewalk

0.11

 siding

0.10

 batter

0.10

 retry

0.10

Activations Density 0.242%