INDEX

Explanations

punctuation followed by a new context

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

ihn

-0.07

ens

-0.07

_push

-0.07

ines

-0.06

æİ¨è¿Ľ

-0.06

æĪĳåĽŃ

-0.06

_PARSE

-0.06

ony

-0.06

ÃŃv

-0.06

po

-0.06

POSITIVE LOGITS

,...

0.18

:...

0.17

...

0.17

...

0.16

(...

0.15

 (...

0.15

 [...]

0.14

 (...)

0.13

?...

0.13

...</

0.13

Activations Density 0.396%