INDEX

Explanations

groups and findings

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

ä¸İä¼Ĺä¸įåĲĮ

-0.23

ä¸Ģä¸ªéĩįè¦ģ

-0.22

äº²çĪ±

-0.19

åı¦ä¸Ģä½į

-0.17

ÐµÐ½ÑĮ

-0.17

å¨ģåĲį

-0.17

çĪ±äºº

-0.17

_EXPORT

-0.17

è¿ĺä¸įæĺ¯

-0.17

âĮĤ

-0.17

POSITIVE LOGITS

 ÐĲÐ»ÐµÐºÑģÐ°Ð½

0.33

 ÐŁÐ¾ÑĩÐµÐ¼

0.28

 ÐĶÐ¼Ð¸ÑĤÑĢ

0.27

 ÐºÐ½Ð¾Ð¿Ðº

0.24

 Ð´Ð°Ð¶

0.24

ä¹ĭéĹ´

0.23

 ÑģÐ»Ð¸ÑĪ

0.23

 Ð¿Ð¾ÑĩÐµÐ¼

0.22

 Ð¿ÑĢÐ¸ÑıÑĤÐ½

0.22

ÐµÑģÑĤÐµÑģÑĤÐ²ÐµÐ½Ð½

0.21

Activations Density 0.207%