INDEX

Explanations

quality, success, or improvement

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

çĶļèĩ³è¿ĺ

-0.22

äºĨä¸įå°ĳ

-0.22

ä¸įå°ĳ

-0.22

 Ð´Ð°Ð¶Ðµ

-0.21

åħ±åĴĮ

-0.21

åºĦæĿĳ

-0.20

é¹°

-0.20

å®¿

-0.20

æĸ¯é¡¿

-0.20

æĸ¼æĺ¯

-0.19

POSITIVE LOGITS

èĢĮå·²

0.29

ç½¢äºĨ

0.27

æŁĲç§į

0.25

åĳĹ

0.23

 certain

0.22

0.21

è¦ģæľī

0.21

sem

0.21

 clue

0.20

Activations Density 0.151%