INDEX

Explanations

already have learned

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

çªģçĦ¶

-0.15

æŃ£å¥½

-0.14

å§ĭç»Ī

-0.13

ç»Īäºİ

-0.12

å§ĭç»ĪåĿļæĮģ

-0.11

 ìĺ¤íŀĪëł¤

-0.11

ä¸ĢçĽ´å¤Ħäºİ

-0.10

èĩ³ä»Ĭ

-0.10

åĪļå¥½

-0.10

æľĢç»Ī

-0.10

POSITIVE LOGITS

å·²ç»ıåľ¨

0.15

å·²ç»ıæĺ¯

0.13

å·²çĦ¶

0.12

å·²ç»ıå¼Ģå§ĭ

0.12

å°±å·²ç»ı

0.11

å·²åľ¨

0.11

 already

0.11

å·²ç»ıæľīäºĨ

0.11

å·²ç»ıæľī

0.11

ä»ĸå·²ç»ı

0.11

Activations Density 0.095%