INDEX

Explanations

perfect tense auxiliaries

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãģķãĤĮãĤĭ

-0.10

 youre

-0.10

ãģıãĤĭ

-0.10

æŃ£åľ¨

-0.10

à¸µà¸Ńà¸¢

-0.09

ëĲĺëĬĶ

-0.09

ãĤīãĤĮãĤĭ

-0.09

 æŃ£

-0.09

ãĤĴãģĻãĤĭ

-0.09

-0.08

POSITIVE LOGITS

 telah

0.47

 ÄĳÃ£

0.44

 haber

0.35

've

0.30

’;ve

0.30

æĽ¾

0.29

 have

0.25

å·²ç»ı

0.24

 yapmÄ±ÅŁ

0.24

å·²

0.23

Activations Density 0.386%