INDEX

Explanations

easily followed by verb

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

 Ð¾Ð±ÑıÐ·Ð°ÑĤÐµÐ»ÑĮÐ½Ð¾

-0.22

 decidedly

-0.22

éįĭ

-0.21

sites

-0.21

èħ±

-0.21

 relentlessly

-0.21

 gotta

-0.21

æ»¡æ»¡çļĦ

-0.20

 mainly

-0.20

 ÑģÐ¾Ð±ÑģÑĤÐ²ÐµÐ½Ð½Ð¾

-0.20

POSITIVE LOGITS

 dÃłng

0.27

çł´

0.26

èĩªå¦Ĥ

0.25

åľ°

0.24

mente

0.24

èĥľä»»

0.23

 extingu

0.22

rix

0.22

çľĭåĩº

0.22

abler

0.21

Activations Density 0.013%