INDEX

Explanations

risk factors

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

Ð¸Ð½Ð´Ð¸Ð²Ð¸Ð´

-0.26

 Ð¿ÑĢÐ¸ÑıÑĤÐ½

-0.25

 Ð·Ð°ÑĢÐµÐ³Ð¸ÑģÑĤ

-0.24

ç¬ĳèĦ¸

-0.24

 Ð¿Ð¾ÑģÐºÐ¾Ð»ÑĮ

-0.23

rium

-0.23

 wÅĤaÅĽ

-0.22

 ÐºÐ½Ð¾Ð¿Ðº

-0.22

Ð¸Ð½Ð´Ð¸Ð²

-0.22

 ÐĲÐ»ÐµÐºÑģÐ°Ð½

-0.21

POSITIVE LOGITS

åĽłç´ł

0.30

çĶŁæ´»æĸ¹å¼ı

0.28

çĶŁæ´»ä¹łæĥ¯

0.27

 factors

0.26

 ÑĩÐ°ÑīÐµ

0.24

å¹´é¾Ħ

0.23

åĳ¨åĽ´

0.23

çļĦåĽłç´ł

0.23

å¹´çºª

0.22

https

0.21

Activations Density 0.172%