INDEX

Explanations

"I" followed by "would/are/me/is"

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

ousel

-0.08

_padding

-0.07

 Isles

-0.07

wski

-0.07

[["

-0.07

olis

-0.07

audi

-0.07

usch

-0.07

uschen

-0.07

_alignment

-0.07

POSITIVE LOGITS

 bowiem

0.09

ä¹Łä¸įä¾ĭå¤ĸ

0.08

ilian

0.07

aker

0.07

 ÐºÐ¾ÑĤÐ¾ÑĢÐ¾Ð³Ð¾

0.07

æĮ¯

0.06

://%

0.06

 donc

0.06

Ð¶Ð¸Ð¹

0.06

å¯¹æŃ¤

0.06

Activations Density 0.696%